[GitHub] [carbondata] MarvinLitt opened a new pull request #3520: [WIP]add spatial-index user guid to doc
Locked
1234
1234
 
|
MarvinLitt commented on a change in pull request #3520: [WIP]add spatial-index user guid to doc
URL: https://github.com/apache/carbondata/pull/3520#discussion_r366335641 ########## File path: docs/spatial-index-guide.md ########## @@ -0,0 +1,94 @@ +<!-- + Licensed to the Apache Software Foundation (ASF) under one or more + contributor license agreements. See the NOTICE file distributed with + this work for additional information regarding copyright ownership. + The ASF licenses this file to you under the Apache License, Version 2.0 + (the "License"); you may not use this file except in compliance with + the License. You may obtain a copy of the License at + + http://www.apache.org/licenses/LICENSE-2.0 + + Unless required by applicable law or agreed to in writing, software + distributed under the License is distributed on an "AS IS" BASIS, + WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + See the License for the specific language governing permissions and + limitations under the License. +--> + +## What is spatial index + +A spatial index is a data structure that allows for accessing a spatial object efficiently. It is a common technique used by spatial databases. Without indexing, any search for a feature would require a "sequential scan" of every record in the database, resulting in much longer processing time. In a spatial index construction process, the minimum bounding rectangle serves as an object approximation. Various types of spatial indices across commercial and open-source databases yield measurable performance differences. Spatial indexing techniques are playing a central role in time-critical applications and the manipulation of spatial big data. + + + +## What does carbondata implement spatial index + +There are many components that implement spatial indexing, like GeoSpark that use GeoMesa format for spatial query. now carbondata implements a different way of spatial index, more like an UDF. Its core is to use grid coordinates to generate coordinate based hash ID, like Z order, it's also regionally continuous. + +CarbonData implements a grid spatial index. It requires that the data has been gridded when it is load into segments. A set of latitude and longitude represents a grid range, the size of the grid can be specified artificially. So the coordinates of the loaded points are often discrete and not continuous. + +The grid and point relationship is like that black point is the middle of a grid, the red dot is just inside the grid. The red point is inside the grid, it can be replaced by the center point of the grid, indicating that the point is within the grid. Therefore, the coordinates of points in a grid are replaced by black points in the middle. This is the characteristic of data load. At the same time of data load, carbondata will generate hash ID according to the coordinates of rows and columns of the grid. These hash IDs are the same as Z order when querying. Detailed conversion algorithm can refer to the design documents of spatial index. + + + +When querying, the user enters the true space polygon coordinates, carbondata use the polygon and spatial region information passed in when creating a table build a quad tree. The nodes in the quad tree are composed of hash ids generated by the row and column information projected in the polygon area and group photo in map area. When the query polygon area is not disjon from the grid center point, the grid is considered selected. In the following figure, user select a quadrilateral polygon, The grid with the center point in the region will generate a quadtree. A list of line with continuous properties will be generated in the query process, like [97->97 99->99 102->102 104->111 120->120 122->123 151->151 157->158 159->159 192->208 210->210 216->216 225->225 228->229], each part of the list represents a continuous grid area. Carbondata use that line list to prune and filtered. About the detail can be search under https://issues.apache.org/jira/browse/CARBONDATA-3548 Review comment: okay ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [hidden email] With regards, Apache Git Services |
|
|
In reply to this post by GitBox
MarvinLitt commented on a change in pull request #3520: [WIP]add spatial-index user guid to doc
URL: https://github.com/apache/carbondata/pull/3520#discussion_r366336529 ########## File path: docs/spatial-index-guide.md ########## @@ -0,0 +1,94 @@ +<!-- + Licensed to the Apache Software Foundation (ASF) under one or more + contributor license agreements. See the NOTICE file distributed with + this work for additional information regarding copyright ownership. + The ASF licenses this file to you under the Apache License, Version 2.0 + (the "License"); you may not use this file except in compliance with + the License. You may obtain a copy of the License at + + http://www.apache.org/licenses/LICENSE-2.0 + + Unless required by applicable law or agreed to in writing, software + distributed under the License is distributed on an "AS IS" BASIS, + WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + See the License for the specific language governing permissions and + limitations under the License. +--> + +## What is spatial index + +A spatial index is a data structure that allows for accessing a spatial object efficiently. It is a common technique used by spatial databases. Without indexing, any search for a feature would require a "sequential scan" of every record in the database, resulting in much longer processing time. In a spatial index construction process, the minimum bounding rectangle serves as an object approximation. Various types of spatial indices across commercial and open-source databases yield measurable performance differences. Spatial indexing techniques are playing a central role in time-critical applications and the manipulation of spatial big data. + + + +## What does carbondata implement spatial index + +There are many components that implement spatial indexing, like GeoSpark that use GeoMesa format for spatial query. now carbondata implements a different way of spatial index, more like an UDF. Its core is to use grid coordinates to generate coordinate based hash ID, like Z order, it's also regionally continuous. + +CarbonData implements a grid spatial index. It requires that the data has been gridded when it is load into segments. A set of latitude and longitude represents a grid range, the size of the grid can be specified artificially. So the coordinates of the loaded points are often discrete and not continuous. + +The grid and point relationship is like that black point is the middle of a grid, the red dot is just inside the grid. The red point is inside the grid, it can be replaced by the center point of the grid, indicating that the point is within the grid. Therefore, the coordinates of points in a grid are replaced by black points in the middle. This is the characteristic of data load. At the same time of data load, carbondata will generate hash ID according to the coordinates of rows and columns of the grid. These hash IDs are the same as Z order when querying. Detailed conversion algorithm can refer to the design documents of spatial index. + + + +When querying, the user enters the true space polygon coordinates, carbondata use the polygon and spatial region information passed in when creating a table build a quad tree. The nodes in the quad tree are composed of hash ids generated by the row and column information projected in the polygon area and group photo in map area. When the query polygon area is not disjon from the grid center point, the grid is considered selected. In the following figure, user select a quadrilateral polygon, The grid with the center point in the region will generate a quadtree. A list of line with continuous properties will be generated in the query process, like [97->97 99->99 102->102 104->111 120->120 122->123 151->151 157->158 159->159 192->208 210->210 216->216 225->225 228->229], each part of the list represents a continuous grid area. Carbondata use that line list to prune and filtered. About the detail can be search under https://issues.apache.org/jira/browse/CARBONDATA-3548 + + + + + +## Installation and Deployment + +Build source with modules geo open, can open "pom.xml" and check whether the mode has been open. + + + +Then you can get the "carbondata-geo-2.0.0-SNAPSHOT.jar" keep this jar and 'jst-core.jar' to your carbonlib path. + +## Basic Command + +### Create Table + +spatial index need to appoint the source column and other regional information. carbon will create a Invisible hash id column. + +example + +``` +create table source_index(id BIGINT, latitude long, longitude long) stored by 'carbondata' TBLPROPERTIES ( +'INDEX_HANDLER'='mygeohash', +'INDEX_HANDLER.mygeohash.type'='geohash', +'INDEX_HANDLER.mygeohash.sourcecolumns'='longitude, latitude', +'INDEX_HANDLER.mygeohash.originLatitude'='19.832277', +'INDEX_HANDLER.mygeohash.gridSize'='50', +'INDEX_HANDLER.mygeohash.minLongitude'='1.811865', +'INDEX_HANDLER.mygeohash.maxLongitude'='2.782233', +'INDEX_HANDLER.mygeohash.minLatitude'='19.832277', +'INDEX_HANDLER.mygeohash.maxLatitude'='20.225281', +'INDEX_HANDLER.mygeohash.conversionRatio'='1000000'); +``` + +| **Property** | **Description** | +| ------------- | :----------------------------------------------------------- | +| INDEX_HANDLER | Custom index handler. This handler allows user to create a new column from the set of schema columns. Newly created column name is same as that of handler name. Type and sourcecolumns properties for the handler are mandatory properties. At present, only supported value for type property is 'geohash'. A simple default implementation class can be available in carbon. User can hook their custom implementation class for geohash by extending the default implementation. An advanced sophisticated custom implementation of geohash can also be made available in carbon. It can expect following additional table properties for handler: 'INDEX_HANDLER.xxx.gridSize', 'INDEX_HANDLER.xxx.minLongitude', 'INDEX_HANDLER.xxx.maxLongitude', 'INDEX_HANDLER.xxx.minLatitude', 'INDEX_HANDLER.xxx.maxLatitude', User can add their own table properties for handler similar to above format and access them in their custom implementation class. All those properties are optional at carbon. Can specify their implementation class with 'INDEX_HANDLER.xxx.class' property Default implementation class can support handler column value generation taking source column values into account for each row and support query filtering based on the source columns. Generated handler column is invisible to user. Column is not allowed in any DDL commands and properties except in SORT_COLUMNS table property. ' conversionRatio' allow user to translate Longitude and Latitude to long. Example: when data loading the real Longitude=13.123456, Latitude=101.12356, configure INDEX_HANDLER.mygeohash.conversionRatio= '1000000', then user can change data to Longitude=13123456, Latitude=10112356. they are long type simpler and faster than floating-point number in calculation. | Review comment: replace with Chart ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [hidden email] With regards, Apache Git Services |
|
|
In reply to this post by GitBox
MarvinLitt commented on a change in pull request #3520: [WIP]add spatial-index user guid to doc
URL: https://github.com/apache/carbondata/pull/3520#discussion_r366336805 ########## File path: docs/spatial-index-guide.md ########## @@ -0,0 +1,94 @@ +<!-- + Licensed to the Apache Software Foundation (ASF) under one or more + contributor license agreements. See the NOTICE file distributed with + this work for additional information regarding copyright ownership. + The ASF licenses this file to you under the Apache License, Version 2.0 + (the "License"); you may not use this file except in compliance with + the License. You may obtain a copy of the License at + + http://www.apache.org/licenses/LICENSE-2.0 + + Unless required by applicable law or agreed to in writing, software + distributed under the License is distributed on an "AS IS" BASIS, + WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + See the License for the specific language governing permissions and + limitations under the License. +--> + +## What is spatial index + +A spatial index is a data structure that allows for accessing a spatial object efficiently. It is a common technique used by spatial databases. Without indexing, any search for a feature would require a "sequential scan" of every record in the database, resulting in much longer processing time. In a spatial index construction process, the minimum bounding rectangle serves as an object approximation. Various types of spatial indices across commercial and open-source databases yield measurable performance differences. Spatial indexing techniques are playing a central role in time-critical applications and the manipulation of spatial big data. + + + +## What does carbondata implement spatial index + +There are many components that implement spatial indexing, like GeoSpark that use GeoMesa format for spatial query. now carbondata implements a different way of spatial index, more like an UDF. Its core is to use grid coordinates to generate coordinate based hash ID, like Z order, it's also regionally continuous. + +CarbonData implements a grid spatial index. It requires that the data has been gridded when it is load into segments. A set of latitude and longitude represents a grid range, the size of the grid can be specified artificially. So the coordinates of the loaded points are often discrete and not continuous. + +The grid and point relationship is like that black point is the middle of a grid, the red dot is just inside the grid. The red point is inside the grid, it can be replaced by the center point of the grid, indicating that the point is within the grid. Therefore, the coordinates of points in a grid are replaced by black points in the middle. This is the characteristic of data load. At the same time of data load, carbondata will generate hash ID according to the coordinates of rows and columns of the grid. These hash IDs are the same as Z order when querying. Detailed conversion algorithm can refer to the design documents of spatial index. + + + +When querying, the user enters the true space polygon coordinates, carbondata use the polygon and spatial region information passed in when creating a table build a quad tree. The nodes in the quad tree are composed of hash ids generated by the row and column information projected in the polygon area and group photo in map area. When the query polygon area is not disjon from the grid center point, the grid is considered selected. In the following figure, user select a quadrilateral polygon, The grid with the center point in the region will generate a quadtree. A list of line with continuous properties will be generated in the query process, like [97->97 99->99 102->102 104->111 120->120 122->123 151->151 157->158 159->159 192->208 210->210 216->216 225->225 228->229], each part of the list represents a continuous grid area. Carbondata use that line list to prune and filtered. About the detail can be search under https://issues.apache.org/jira/browse/CARBONDATA-3548 + + + + + +## Installation and Deployment + +Build source with modules geo open, can open "pom.xml" and check whether the mode has been open. + + + +Then you can get the "carbondata-geo-2.0.0-SNAPSHOT.jar" keep this jar and 'jst-core.jar' to your carbonlib path. + +## Basic Command + +### Create Table + +spatial index need to appoint the source column and other regional information. carbon will create a Invisible hash id column. + +example + +``` +create table source_index(id BIGINT, latitude long, longitude long) stored by 'carbondata' TBLPROPERTIES ( +'INDEX_HANDLER'='mygeohash', +'INDEX_HANDLER.mygeohash.type'='geohash', +'INDEX_HANDLER.mygeohash.sourcecolumns'='longitude, latitude', +'INDEX_HANDLER.mygeohash.originLatitude'='19.832277', +'INDEX_HANDLER.mygeohash.gridSize'='50', +'INDEX_HANDLER.mygeohash.minLongitude'='1.811865', +'INDEX_HANDLER.mygeohash.maxLongitude'='2.782233', +'INDEX_HANDLER.mygeohash.minLatitude'='19.832277', +'INDEX_HANDLER.mygeohash.maxLatitude'='20.225281', +'INDEX_HANDLER.mygeohash.conversionRatio'='1000000'); +``` + +| **Property** | **Description** | +| ------------- | :----------------------------------------------------------- | +| INDEX_HANDLER | Custom index handler. This handler allows user to create a new column from the set of schema columns. Newly created column name is same as that of handler name. Type and sourcecolumns properties for the handler are mandatory properties. At present, only supported value for type property is 'geohash'. A simple default implementation class can be available in carbon. User can hook their custom implementation class for geohash by extending the default implementation. An advanced sophisticated custom implementation of geohash can also be made available in carbon. It can expect following additional table properties for handler: 'INDEX_HANDLER.xxx.gridSize', 'INDEX_HANDLER.xxx.minLongitude', 'INDEX_HANDLER.xxx.maxLongitude', 'INDEX_HANDLER.xxx.minLatitude', 'INDEX_HANDLER.xxx.maxLatitude', User can add their own table properties for handler similar to above format and access them in their custom implementation class. All those properties are optional at carbon. Can specify their implementation class with 'INDEX_HANDLER.xxx.class' property Default implementation class can support handler column value generation taking source column values into account for each row and support query filtering based on the source columns. Generated handler column is invisible to user. Column is not allowed in any DDL commands and properties except in SORT_COLUMNS table property. ' conversionRatio' allow user to translate Longitude and Latitude to long. Example: when data loading the real Longitude=13.123456, Latitude=101.12356, configure INDEX_HANDLER.mygeohash.conversionRatio= '1000000', then user can change data to Longitude=13123456, Latitude=10112356. they are long type simpler and faster than floating-point number in calculation. | Review comment: okay ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [hidden email] With regards, Apache Git Services |
|
|
In reply to this post by GitBox
MarvinLitt commented on a change in pull request #3520: [WIP]add spatial-index user guid to doc
URL: https://github.com/apache/carbondata/pull/3520#discussion_r366336889 ########## File path: docs/spatial-index-guide.md ########## @@ -0,0 +1,94 @@ +<!-- + Licensed to the Apache Software Foundation (ASF) under one or more + contributor license agreements. See the NOTICE file distributed with + this work for additional information regarding copyright ownership. + The ASF licenses this file to you under the Apache License, Version 2.0 + (the "License"); you may not use this file except in compliance with + the License. You may obtain a copy of the License at + + http://www.apache.org/licenses/LICENSE-2.0 + + Unless required by applicable law or agreed to in writing, software + distributed under the License is distributed on an "AS IS" BASIS, + WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + See the License for the specific language governing permissions and + limitations under the License. +--> + +## What is spatial index + +A spatial index is a data structure that allows for accessing a spatial object efficiently. It is a common technique used by spatial databases. Without indexing, any search for a feature would require a "sequential scan" of every record in the database, resulting in much longer processing time. In a spatial index construction process, the minimum bounding rectangle serves as an object approximation. Various types of spatial indices across commercial and open-source databases yield measurable performance differences. Spatial indexing techniques are playing a central role in time-critical applications and the manipulation of spatial big data. + + + +## What does carbondata implement spatial index + +There are many components that implement spatial indexing, like GeoSpark that use GeoMesa format for spatial query. now carbondata implements a different way of spatial index, more like an UDF. Its core is to use grid coordinates to generate coordinate based hash ID, like Z order, it's also regionally continuous. + +CarbonData implements a grid spatial index. It requires that the data has been gridded when it is load into segments. A set of latitude and longitude represents a grid range, the size of the grid can be specified artificially. So the coordinates of the loaded points are often discrete and not continuous. + +The grid and point relationship is like that black point is the middle of a grid, the red dot is just inside the grid. The red point is inside the grid, it can be replaced by the center point of the grid, indicating that the point is within the grid. Therefore, the coordinates of points in a grid are replaced by black points in the middle. This is the characteristic of data load. At the same time of data load, carbondata will generate hash ID according to the coordinates of rows and columns of the grid. These hash IDs are the same as Z order when querying. Detailed conversion algorithm can refer to the design documents of spatial index. + + + +When querying, the user enters the true space polygon coordinates, carbondata use the polygon and spatial region information passed in when creating a table build a quad tree. The nodes in the quad tree are composed of hash ids generated by the row and column information projected in the polygon area and group photo in map area. When the query polygon area is not disjon from the grid center point, the grid is considered selected. In the following figure, user select a quadrilateral polygon, The grid with the center point in the region will generate a quadtree. A list of line with continuous properties will be generated in the query process, like [97->97 99->99 102->102 104->111 120->120 122->123 151->151 157->158 159->159 192->208 210->210 216->216 225->225 228->229], each part of the list represents a continuous grid area. Carbondata use that line list to prune and filtered. About the detail can be search under https://issues.apache.org/jira/browse/CARBONDATA-3548 + + + + + +## Installation and Deployment + +Build source with modules geo open, can open "pom.xml" and check whether the mode has been open. + + + +Then you can get the "carbondata-geo-2.0.0-SNAPSHOT.jar" keep this jar and 'jst-core.jar' to your carbonlib path. + +## Basic Command + +### Create Table + +spatial index need to appoint the source column and other regional information. carbon will create a Invisible hash id column. + +example + +``` +create table source_index(id BIGINT, latitude long, longitude long) stored by 'carbondata' TBLPROPERTIES ( +'INDEX_HANDLER'='mygeohash', +'INDEX_HANDLER.mygeohash.type'='geohash', +'INDEX_HANDLER.mygeohash.sourcecolumns'='longitude, latitude', +'INDEX_HANDLER.mygeohash.originLatitude'='19.832277', +'INDEX_HANDLER.mygeohash.gridSize'='50', +'INDEX_HANDLER.mygeohash.minLongitude'='1.811865', +'INDEX_HANDLER.mygeohash.maxLongitude'='2.782233', +'INDEX_HANDLER.mygeohash.minLatitude'='19.832277', +'INDEX_HANDLER.mygeohash.maxLatitude'='20.225281', +'INDEX_HANDLER.mygeohash.conversionRatio'='1000000'); +``` + +| **Property** | **Description** | +| ------------- | :----------------------------------------------------------- | +| INDEX_HANDLER | Custom index handler. This handler allows user to create a new column from the set of schema columns. Newly created column name is same as that of handler name. Type and sourcecolumns properties for the handler are mandatory properties. At present, only supported value for type property is 'geohash'. A simple default implementation class can be available in carbon. User can hook their custom implementation class for geohash by extending the default implementation. An advanced sophisticated custom implementation of geohash can also be made available in carbon. It can expect following additional table properties for handler: 'INDEX_HANDLER.xxx.gridSize', 'INDEX_HANDLER.xxx.minLongitude', 'INDEX_HANDLER.xxx.maxLongitude', 'INDEX_HANDLER.xxx.minLatitude', 'INDEX_HANDLER.xxx.maxLatitude', User can add their own table properties for handler similar to above format and access them in their custom implementation class. All those properties are optional at carbon. Can specify their implementation class with 'INDEX_HANDLER.xxx.class' property Default implementation class can support handler column value generation taking source column values into account for each row and support query filtering based on the source columns. Generated handler column is invisible to user. Column is not allowed in any DDL commands and properties except in SORT_COLUMNS table property. ' conversionRatio' allow user to translate Longitude and Latitude to long. Example: when data loading the real Longitude=13.123456, Latitude=101.12356, configure INDEX_HANDLER.mygeohash.conversionRatio= '1000000', then user can change data to Longitude=13123456, Latitude=10112356. they are long type simpler and faster than floating-point number in calculation. | + + + +### Data Loading + +In this process, tables with spatial indexes configured will create invisible hash ids column data based on 'sourcecolumns' user defined and store it. Please ensure that the specified sourcecolumns column has data. + +### Data query + +CarbonData implements a UDF named 'IN_POLYGON'. + +example: + +``` +select * from source_index where IN_POLYGON('16.321011 4.123503,16.137676 5.947911,16.560993 5.935276,16.321011 4.123503' Review comment: oaky ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [hidden email] With regards, Apache Git Services |
|
|
In reply to this post by GitBox
MarvinLitt commented on a change in pull request #3520: [WIP]add spatial-index user guid to doc
URL: https://github.com/apache/carbondata/pull/3520#discussion_r366338684 ########## File path: docs/spatial-index-guide.md ########## @@ -0,0 +1,94 @@ +<!-- + Licensed to the Apache Software Foundation (ASF) under one or more + contributor license agreements. See the NOTICE file distributed with + this work for additional information regarding copyright ownership. + The ASF licenses this file to you under the Apache License, Version 2.0 + (the "License"); you may not use this file except in compliance with + the License. You may obtain a copy of the License at + + http://www.apache.org/licenses/LICENSE-2.0 + + Unless required by applicable law or agreed to in writing, software + distributed under the License is distributed on an "AS IS" BASIS, + WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + See the License for the specific language governing permissions and + limitations under the License. +--> + +## What is spatial index + +A spatial index is a data structure that allows for accessing a spatial object efficiently. It is a common technique used by spatial databases. Without indexing, any search for a feature would require a "sequential scan" of every record in the database, resulting in much longer processing time. In a spatial index construction process, the minimum bounding rectangle serves as an object approximation. Various types of spatial indices across commercial and open-source databases yield measurable performance differences. Spatial indexing techniques are playing a central role in time-critical applications and the manipulation of spatial big data. + + + +## What does carbondata implement spatial index + +There are many components that implement spatial indexing, like GeoSpark that use GeoMesa format for spatial query. now carbondata implements a different way of spatial index, more like an UDF. Its core is to use grid coordinates to generate coordinate based hash ID, like Z order, it's also regionally continuous. Review comment: thank you for your help. has replace the words. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [hidden email] With regards, Apache Git Services |
|
|
In reply to this post by GitBox
CarbonDataQA1 commented on issue #3520: [WIP]add spatial-index user guid to doc
URL: https://github.com/apache/carbondata/pull/3520#issuecomment-574235898 Build Success with Spark 2.3.4, Please check CI http://121.244.95.60:12545/job/ApacheCarbonPRBuilder2.3/1639/ ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [hidden email] With regards, Apache Git Services |
|
|
In reply to this post by GitBox
kunal642 commented on a change in pull request #3520: [CARBONDATA-3548]add spatial-index user guid to doc
URL: https://github.com/apache/carbondata/pull/3520#discussion_r370071006 ########## File path: docs/spatial-index-guide.md ########## @@ -0,0 +1,109 @@ +<!-- + Licensed to the Apache Software Foundation (ASF) under one or more + contributor license agreements. See the NOTICE file distributed with + this work for additional information regarding copyright ownership. + The ASF licenses this file to you under the Apache License, Version 2.0 + (the "License"); you may not use this file except in compliance with + the License. You may obtain a copy of the License at + + http://www.apache.org/licenses/LICENSE-2.0 + + Unless required by applicable law or agreed to in writing, software + distributed under the License is distributed on an "AS IS" BASIS, + WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + See the License for the specific language governing permissions and + limitations under the License. +--> + +# What is spatial index + +[A spatial index](https://gistbok.ucgis.org/topic-keywords/indexing) is a data structure that allows for accessing a spatial object efficiently. It is a common technique used by spatial databases. Without indexing, any search for a feature would require a "sequential scan" of every record in the database, resulting in much longer processing time. In a spatial index construction process, the minimum bounding rectangle serves as an object approximation. Various types of spatial indices across commercial and open-source databases yield measurable performance differences. Spatial indexing techniques are playing a central role in time-critical applications and the manipulation of spatial big data. + + + +# How does carbondata implement spatial index + +There are many opensource implementations for spatial indexing and to process spatial queries. CarbonData implements a different way of spatial index. Its core idea is to use the raster data. Raster is made up of matrix of cells organized into rows and columns(called a grid). Each cell represents a coordinate. And the index for that coodrinate is generated using longitude and latitude, like the [Z order curve](https://en.wikipedia.org/wiki/Z-order_curve). + +CarbonData rasterize the user data during data load into segments. A set of latitude and longitude represents a grid range. The size of the grid can be configured. Hence, the coordinates loaded are often discrete and not continuous. + +Below figure shows the relationship between the grid and the points residing in it. Black point represents the center point of the grid, and the red points are the coordinates at the arbitrary positions inside the grid. The red points can be replaced by the center point of the grid to indicate that the points lies within the grid. During data load, CarbonData generates an Index for coordinate according to row and column of the grid(in the raster) where that coordinate lies. These Indexes are the same as Z order. For the detailed conversion algorithm, please refer to the design documents of spatial index. + + + +When querying, the user enters the true space polygon coordinates, carbondata use the polygon and spatial region information passed in when creating a table build a quad tree. The nodes in the quad tree are composed of hash ids generated by the row and column information projected in the polygon area. When the query polygon area is not disjoint from the grid center point, the grid is considered selected. In the following figure, user select a quadrilateral polygon, The grid with the center point in the region will generate a quadtree. A list of line with continuous properties will be generated in the query process, like [97->97 99->99 102->102 104->111 120->120 122->123 151->151 157->158 159->159 192->208 210->210 216->216 225->225 228->229], each part of the list represents a continuous grid area. Carbondata use that line list to prune and filtered. About the detail can be search under https://issues.apache.org/jira/browse/CARBONDATA-3548 + + + + + +# Installation and Deployment + +Build source with modules geo open, can open "pom.xml" and check whether the mode has been open. + + Review comment: spatial-index-3.png is a useless image. Please do no add images showing pom changes ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [hidden email] With regards, Apache Git Services |
|
|
In reply to this post by GitBox
kunal642 commented on a change in pull request #3520: [CARBONDATA-3548]add spatial-index user guid to doc
URL: https://github.com/apache/carbondata/pull/3520#discussion_r370071417 ########## File path: docs/spatial-index-guide.md ########## @@ -0,0 +1,109 @@ +<!-- + Licensed to the Apache Software Foundation (ASF) under one or more + contributor license agreements. See the NOTICE file distributed with + this work for additional information regarding copyright ownership. + The ASF licenses this file to you under the Apache License, Version 2.0 + (the "License"); you may not use this file except in compliance with + the License. You may obtain a copy of the License at + + http://www.apache.org/licenses/LICENSE-2.0 + + Unless required by applicable law or agreed to in writing, software + distributed under the License is distributed on an "AS IS" BASIS, + WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + See the License for the specific language governing permissions and + limitations under the License. +--> + +# What is spatial index + +[A spatial index](https://gistbok.ucgis.org/topic-keywords/indexing) is a data structure that allows for accessing a spatial object efficiently. It is a common technique used by spatial databases. Without indexing, any search for a feature would require a "sequential scan" of every record in the database, resulting in much longer processing time. In a spatial index construction process, the minimum bounding rectangle serves as an object approximation. Various types of spatial indices across commercial and open-source databases yield measurable performance differences. Spatial indexing techniques are playing a central role in time-critical applications and the manipulation of spatial big data. + + + +# How does carbondata implement spatial index + +There are many opensource implementations for spatial indexing and to process spatial queries. CarbonData implements a different way of spatial index. Its core idea is to use the raster data. Raster is made up of matrix of cells organized into rows and columns(called a grid). Each cell represents a coordinate. And the index for that coodrinate is generated using longitude and latitude, like the [Z order curve](https://en.wikipedia.org/wiki/Z-order_curve). + +CarbonData rasterize the user data during data load into segments. A set of latitude and longitude represents a grid range. The size of the grid can be configured. Hence, the coordinates loaded are often discrete and not continuous. + +Below figure shows the relationship between the grid and the points residing in it. Black point represents the center point of the grid, and the red points are the coordinates at the arbitrary positions inside the grid. The red points can be replaced by the center point of the grid to indicate that the points lies within the grid. During data load, CarbonData generates an Index for coordinate according to row and column of the grid(in the raster) where that coordinate lies. These Indexes are the same as Z order. For the detailed conversion algorithm, please refer to the design documents of spatial index. + + + +When querying, the user enters the true space polygon coordinates, carbondata use the polygon and spatial region information passed in when creating a table build a quad tree. The nodes in the quad tree are composed of hash ids generated by the row and column information projected in the polygon area. When the query polygon area is not disjoint from the grid center point, the grid is considered selected. In the following figure, user select a quadrilateral polygon, The grid with the center point in the region will generate a quadtree. A list of line with continuous properties will be generated in the query process, like [97->97 99->99 102->102 104->111 120->120 122->123 151->151 157->158 159->159 192->208 210->210 216->216 225->225 228->229], each part of the list represents a continuous grid area. Carbondata use that line list to prune and filtered. About the detail can be search under https://issues.apache.org/jira/browse/CARBONDATA-3548 + + + + + +# Installation and Deployment Review comment: The description is not clear for this. Please mention the mvn command instead of the description. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [hidden email] With regards, Apache Git Services |
|
|
In reply to this post by GitBox
CarbonDataQA1 commented on issue #3520: [CARBONDATA-3548]add spatial-index user guid to doc
URL: https://github.com/apache/carbondata/pull/3520#issuecomment-581264465 Build Failed with Spark 2.3.4, Please check CI http://121.244.95.60:12545/job/ApacheCarbon_PR_Builder_2.4.4/12/ ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [hidden email] With regards, Apache Git Services |
|
|
In reply to this post by GitBox
ajantha-bhat commented on issue #3520: [CARBONDATA-3548]add spatial-index user guid to doc
URL: https://github.com/apache/carbondata/pull/3520#issuecomment-590172136 please add a link to tets case ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [hidden email] With regards, Apache Git Services |
|
|
In reply to this post by GitBox
ajantha-bhat commented on issue #3520: [CARBONDATA-3548]add spatial-index user guid to doc
URL: https://github.com/apache/carbondata/pull/3520#issuecomment-604842289 @MarvinLitt : please handle kunal's comment. I want to merge this. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [hidden email] With regards, Apache Git Services |
|
|
In reply to this post by GitBox
ajantha-bhat commented on issue #3520: [CARBONDATA-3548]add spatial-index user guid to doc
URL: https://github.com/apache/carbondata/pull/3520#issuecomment-604842332 retest this please ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [hidden email] With regards, Apache Git Services |
|
|
In reply to this post by GitBox
ajantha-bhat commented on a change in pull request #3520: [CARBONDATA-3548]add spatial-index user guid to doc
URL: https://github.com/apache/carbondata/pull/3520#discussion_r399066845 ########## File path: docs/spatial-index-guide.md ########## @@ -0,0 +1,109 @@ +<!-- + Licensed to the Apache Software Foundation (ASF) under one or more + contributor license agreements. See the NOTICE file distributed with + this work for additional information regarding copyright ownership. + The ASF licenses this file to you under the Apache License, Version 2.0 + (the "License"); you may not use this file except in compliance with + the License. You may obtain a copy of the License at + + http://www.apache.org/licenses/LICENSE-2.0 + + Unless required by applicable law or agreed to in writing, software + distributed under the License is distributed on an "AS IS" BASIS, + WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + See the License for the specific language governing permissions and + limitations under the License. +--> + +# What is spatial index + +[A spatial index](https://gistbok.ucgis.org/topic-keywords/indexing) is a data structure that allows for accessing a spatial object efficiently. It is a common technique used by spatial databases. Without indexing, any search for a feature would require a "sequential scan" of every record in the database, resulting in much longer processing time. In a spatial index construction process, the minimum bounding rectangle serves as an object approximation. Various types of spatial indices across commercial and open-source databases yield measurable performance differences. Spatial indexing techniques are playing a central role in time-critical applications and the manipulation of spatial big data. + + + +# How does carbondata implement spatial index + +There are many opensource implementations for spatial indexing and to process spatial queries. CarbonData implements a different way of spatial index. Its core idea is to use the raster data. Raster is made up of matrix of cells organized into rows and columns(called a grid). Each cell represents a coordinate. And the index for that coodrinate is generated using longitude and latitude, like the [Z order curve](https://en.wikipedia.org/wiki/Z-order_curve). + +CarbonData rasterize the user data during data load into segments. A set of latitude and longitude represents a grid range. The size of the grid can be configured. Hence, the coordinates loaded are often discrete and not continuous. + +Below figure shows the relationship between the grid and the points residing in it. Black point represents the center point of the grid, and the red points are the coordinates at the arbitrary positions inside the grid. The red points can be replaced by the center point of the grid to indicate that the points lies within the grid. During data load, CarbonData generates an Index for coordinate according to row and column of the grid(in the raster) where that coordinate lies. These Indexes are the same as Z order. For the detailed conversion algorithm, please refer to the design documents of spatial index. + + + +When querying, the user enters the true space polygon coordinates, carbondata use the polygon and spatial region information passed in when creating a table build a quad tree. The nodes in the quad tree are composed of hash ids generated by the row and column information projected in the polygon area. When the query polygon area is not disjoint from the grid center point, the grid is considered selected. In the following figure, user select a quadrilateral polygon, The grid with the center point in the region will generate a quadtree. A list of line with continuous properties will be generated in the query process, like [97->97 99->99 102->102 104->111 120->120 122->123 151->151 157->158 159->159 192->208 210->210 216->216 225->225 228->229], each part of the list represents a continuous grid area. Carbondata use that line list to prune and filtered. About the detail can be search under https://issues.apache.org/jira/browse/CARBONDATA-3548 + + + + + +# Installation and Deployment + +Build source with modules geo open, can open "pom.xml" and check whether the mode has been open. + + + +Then you can get the "carbondata-geo-2.0.0-SNAPSHOT.jar" keep this jar and 'jst-core.jar' to your carbonlib path. + +## Basic Command + +### Create Table + +spatial index need to appoint the source column and other regional information. carbon will create a Invisible hash id column. + +example + +```sql +create table source_index(id BIGINT, latitude long, longitude long) stored by 'carbondata' TBLPROPERTIES ( +'INDEX_HANDLER'='mygeohash', +'INDEX_HANDLER.mygeohash.type'='geohash', +'INDEX_HANDLER.mygeohash.sourcecolumns'='longitude, latitude', +'INDEX_HANDLER.mygeohash.originLatitude'='19.832277', +'INDEX_HANDLER.mygeohash.gridSize'='50', +'INDEX_HANDLER.mygeohash.minLongitude'='1.811865', +'INDEX_HANDLER.mygeohash.maxLongitude'='2.782233', +'INDEX_HANDLER.mygeohash.minLatitude'='19.832277', +'INDEX_HANDLER.mygeohash.maxLatitude'='20.225281', +'INDEX_HANDLER.mygeohash.conversionRatio'='1000000'); Review comment: add for INDEX_HANDLER.xxx.class also ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [hidden email] With regards, Apache Git Services |
|
|
In reply to this post by GitBox
ajantha-bhat commented on a change in pull request #3520: [CARBONDATA-3548]add spatial-index user guid to doc
URL: https://github.com/apache/carbondata/pull/3520#discussion_r399066962 ########## File path: docs/spatial-index-guide.md ########## @@ -0,0 +1,109 @@ +<!-- + Licensed to the Apache Software Foundation (ASF) under one or more + contributor license agreements. See the NOTICE file distributed with + this work for additional information regarding copyright ownership. + The ASF licenses this file to you under the Apache License, Version 2.0 + (the "License"); you may not use this file except in compliance with + the License. You may obtain a copy of the License at + + http://www.apache.org/licenses/LICENSE-2.0 + + Unless required by applicable law or agreed to in writing, software + distributed under the License is distributed on an "AS IS" BASIS, + WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + See the License for the specific language governing permissions and + limitations under the License. +--> + +# What is spatial index + +[A spatial index](https://gistbok.ucgis.org/topic-keywords/indexing) is a data structure that allows for accessing a spatial object efficiently. It is a common technique used by spatial databases. Without indexing, any search for a feature would require a "sequential scan" of every record in the database, resulting in much longer processing time. In a spatial index construction process, the minimum bounding rectangle serves as an object approximation. Various types of spatial indices across commercial and open-source databases yield measurable performance differences. Spatial indexing techniques are playing a central role in time-critical applications and the manipulation of spatial big data. + + + +# How does carbondata implement spatial index + +There are many opensource implementations for spatial indexing and to process spatial queries. CarbonData implements a different way of spatial index. Its core idea is to use the raster data. Raster is made up of matrix of cells organized into rows and columns(called a grid). Each cell represents a coordinate. And the index for that coodrinate is generated using longitude and latitude, like the [Z order curve](https://en.wikipedia.org/wiki/Z-order_curve). + +CarbonData rasterize the user data during data load into segments. A set of latitude and longitude represents a grid range. The size of the grid can be configured. Hence, the coordinates loaded are often discrete and not continuous. + +Below figure shows the relationship between the grid and the points residing in it. Black point represents the center point of the grid, and the red points are the coordinates at the arbitrary positions inside the grid. The red points can be replaced by the center point of the grid to indicate that the points lies within the grid. During data load, CarbonData generates an Index for coordinate according to row and column of the grid(in the raster) where that coordinate lies. These Indexes are the same as Z order. For the detailed conversion algorithm, please refer to the design documents of spatial index. + + + +When querying, the user enters the true space polygon coordinates, carbondata use the polygon and spatial region information passed in when creating a table build a quad tree. The nodes in the quad tree are composed of hash ids generated by the row and column information projected in the polygon area. When the query polygon area is not disjoint from the grid center point, the grid is considered selected. In the following figure, user select a quadrilateral polygon, The grid with the center point in the region will generate a quadtree. A list of line with continuous properties will be generated in the query process, like [97->97 99->99 102->102 104->111 120->120 122->123 151->151 157->158 159->159 192->208 210->210 216->216 225->225 228->229], each part of the list represents a continuous grid area. Carbondata use that line list to prune and filtered. About the detail can be search under https://issues.apache.org/jira/browse/CARBONDATA-3548 + + + + + +# Installation and Deployment + +Build source with modules geo open, can open "pom.xml" and check whether the mode has been open. + + + +Then you can get the "carbondata-geo-2.0.0-SNAPSHOT.jar" keep this jar and 'jst-core.jar' to your carbonlib path. + +## Basic Command + +### Create Table + +spatial index need to appoint the source column and other regional information. carbon will create a Invisible hash id column. + +example + +```sql +create table source_index(id BIGINT, latitude long, longitude long) stored by 'carbondata' TBLPROPERTIES ( +'INDEX_HANDLER'='mygeohash', +'INDEX_HANDLER.mygeohash.type'='geohash', +'INDEX_HANDLER.mygeohash.sourcecolumns'='longitude, latitude', +'INDEX_HANDLER.mygeohash.originLatitude'='19.832277', +'INDEX_HANDLER.mygeohash.gridSize'='50', +'INDEX_HANDLER.mygeohash.minLongitude'='1.811865', +'INDEX_HANDLER.mygeohash.maxLongitude'='2.782233', +'INDEX_HANDLER.mygeohash.minLatitude'='19.832277', +'INDEX_HANDLER.mygeohash.maxLatitude'='20.225281', +'INDEX_HANDLER.mygeohash.conversionRatio'='1000000'); +``` + +| Name | Value | Describe | Review comment: remove `value` column. No need here ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [hidden email] With regards, Apache Git Services |
|
|
In reply to this post by GitBox
ajantha-bhat commented on a change in pull request #3520: [CARBONDATA-3548]add spatial-index user guid to doc
URL: https://github.com/apache/carbondata/pull/3520#discussion_r399068116 ########## File path: docs/spatial-index-guide.md ########## @@ -0,0 +1,109 @@ +<!-- + Licensed to the Apache Software Foundation (ASF) under one or more + contributor license agreements. See the NOTICE file distributed with + this work for additional information regarding copyright ownership. + The ASF licenses this file to you under the Apache License, Version 2.0 + (the "License"); you may not use this file except in compliance with + the License. You may obtain a copy of the License at + + http://www.apache.org/licenses/LICENSE-2.0 + + Unless required by applicable law or agreed to in writing, software + distributed under the License is distributed on an "AS IS" BASIS, + WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + See the License for the specific language governing permissions and + limitations under the License. +--> + +# What is spatial index + +[A spatial index](https://gistbok.ucgis.org/topic-keywords/indexing) is a data structure that allows for accessing a spatial object efficiently. It is a common technique used by spatial databases. Without indexing, any search for a feature would require a "sequential scan" of every record in the database, resulting in much longer processing time. In a spatial index construction process, the minimum bounding rectangle serves as an object approximation. Various types of spatial indices across commercial and open-source databases yield measurable performance differences. Spatial indexing techniques are playing a central role in time-critical applications and the manipulation of spatial big data. + + + +# How does carbondata implement spatial index + +There are many opensource implementations for spatial indexing and to process spatial queries. CarbonData implements a different way of spatial index. Its core idea is to use the raster data. Raster is made up of matrix of cells organized into rows and columns(called a grid). Each cell represents a coordinate. And the index for that coodrinate is generated using longitude and latitude, like the [Z order curve](https://en.wikipedia.org/wiki/Z-order_curve). + +CarbonData rasterize the user data during data load into segments. A set of latitude and longitude represents a grid range. The size of the grid can be configured. Hence, the coordinates loaded are often discrete and not continuous. + +Below figure shows the relationship between the grid and the points residing in it. Black point represents the center point of the grid, and the red points are the coordinates at the arbitrary positions inside the grid. The red points can be replaced by the center point of the grid to indicate that the points lies within the grid. During data load, CarbonData generates an Index for coordinate according to row and column of the grid(in the raster) where that coordinate lies. These Indexes are the same as Z order. For the detailed conversion algorithm, please refer to the design documents of spatial index. + + + +When querying, the user enters the true space polygon coordinates, carbondata use the polygon and spatial region information passed in when creating a table build a quad tree. The nodes in the quad tree are composed of hash ids generated by the row and column information projected in the polygon area. When the query polygon area is not disjoint from the grid center point, the grid is considered selected. In the following figure, user select a quadrilateral polygon, The grid with the center point in the region will generate a quadtree. A list of line with continuous properties will be generated in the query process, like [97->97 99->99 102->102 104->111 120->120 122->123 151->151 157->158 159->159 192->208 210->210 216->216 225->225 228->229], each part of the list represents a continuous grid area. Carbondata use that line list to prune and filtered. About the detail can be search under https://issues.apache.org/jira/browse/CARBONDATA-3548 Review comment: Mention below lines: The main reason for faster spatial queries in carbon is because of polygon filter will be pushed down to the carbon layer and carbon will scan only the matched blocklets instead of full scan. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [hidden email] With regards, Apache Git Services |
|
|
In reply to this post by GitBox
ajantha-bhat edited a comment on issue #3520: [CARBONDATA-3548]add spatial-index user guid to doc
URL: https://github.com/apache/carbondata/pull/3520#issuecomment-604842289 @MarvinLitt : please handle comment. I want to merge this PR. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [hidden email] With regards, Apache Git Services |
|
|
In reply to this post by GitBox
CarbonDataQA1 commented on issue #3520: [CARBONDATA-3548]add spatial-index user guid to doc
URL: https://github.com/apache/carbondata/pull/3520#issuecomment-604892466 Build Success with Spark 2.4.4, Please check CI http://121.244.95.60:12545/job/ApacheCarbon_PR_Builder_2.4.5/862/ ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [hidden email] With regards, Apache Git Services |
|
|
In reply to this post by GitBox
CarbonDataQA1 commented on issue #3520: [CARBONDATA-3548]add spatial-index user guid to doc
URL: https://github.com/apache/carbondata/pull/3520#issuecomment-604894818 Build Success with Spark 2.3.4, Please check CI http://121.244.95.60:12545/job/ApacheCarbonPRBuilder2.3/2570/ ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [hidden email] With regards, Apache Git Services |
|
|
In reply to this post by GitBox
VenuReddy2103 commented on a change in pull request #3520: [CARBONDATA-3548]add spatial-index user guid to doc
URL: https://github.com/apache/carbondata/pull/3520#discussion_r399398817 ########## File path: docs/spatial-index-guide.md ########## @@ -0,0 +1,109 @@ +<!-- + Licensed to the Apache Software Foundation (ASF) under one or more + contributor license agreements. See the NOTICE file distributed with + this work for additional information regarding copyright ownership. + The ASF licenses this file to you under the Apache License, Version 2.0 + (the "License"); you may not use this file except in compliance with + the License. You may obtain a copy of the License at + + http://www.apache.org/licenses/LICENSE-2.0 + + Unless required by applicable law or agreed to in writing, software + distributed under the License is distributed on an "AS IS" BASIS, + WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + See the License for the specific language governing permissions and + limitations under the License. +--> + +# What is spatial index + +[A spatial index](https://gistbok.ucgis.org/topic-keywords/indexing) is a data structure that allows for accessing a spatial object efficiently. It is a common technique used by spatial databases. Without indexing, any search for a feature would require a "sequential scan" of every record in the database, resulting in much longer processing time. In a spatial index construction process, the minimum bounding rectangle serves as an object approximation. Various types of spatial indices across commercial and open-source databases yield measurable performance differences. Spatial indexing techniques are playing a central role in time-critical applications and the manipulation of spatial big data. + + + +# How does carbondata implement spatial index + +There are many opensource implementations for spatial indexing and to process spatial queries. CarbonData implements a different way of spatial index. Its core idea is to use the raster data. Raster is made up of matrix of cells organized into rows and columns(called a grid). Each cell represents a coordinate. And the index for that coodrinate is generated using longitude and latitude, like the [Z order curve](https://en.wikipedia.org/wiki/Z-order_curve). + +CarbonData rasterize the user data during data load into segments. A set of latitude and longitude represents a grid range. The size of the grid can be configured. Hence, the coordinates loaded are often discrete and not continuous. + +Below figure shows the relationship between the grid and the points residing in it. Black point represents the center point of the grid, and the red points are the coordinates at the arbitrary positions inside the grid. The red points can be replaced by the center point of the grid to indicate that the points lies within the grid. During data load, CarbonData generates an Index for coordinate according to row and column of the grid(in the raster) where that coordinate lies. These Indexes are the same as Z order. For the detailed conversion algorithm, please refer to the design documents of spatial index. + + + +When querying, the user enters the true space polygon coordinates, carbondata use the polygon and spatial region information passed in when creating a table build a quad tree. The nodes in the quad tree are composed of hash ids generated by the row and column information projected in the polygon area. When the query polygon area is not disjoint from the grid center point, the grid is considered selected. In the following figure, user select a quadrilateral polygon, The grid with the center point in the region will generate a quadtree. A list of line with continuous properties will be generated in the query process, like [97->97 99->99 102->102 104->111 120->120 122->123 151->151 157->158 159->159 192->208 210->210 216->216 225->225 228->229], each part of the list represents a continuous grid area. Carbondata use that line list to prune and filtered. About the detail can be search under https://issues.apache.org/jira/browse/CARBONDATA-3548 + + + + + +# Installation and Deployment + +Build source with modules geo open, can open "pom.xml" and check whether the mode has been open. + + Review comment: Modified ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [hidden email] With regards, Apache Git Services |
|
|
In reply to this post by GitBox
VenuReddy2103 commented on a change in pull request #3520: [CARBONDATA-3548]add spatial-index user guid to doc
URL: https://github.com/apache/carbondata/pull/3520#discussion_r399399033 ########## File path: docs/spatial-index-guide.md ########## @@ -0,0 +1,109 @@ +<!-- + Licensed to the Apache Software Foundation (ASF) under one or more + contributor license agreements. See the NOTICE file distributed with + this work for additional information regarding copyright ownership. + The ASF licenses this file to you under the Apache License, Version 2.0 + (the "License"); you may not use this file except in compliance with + the License. You may obtain a copy of the License at + + http://www.apache.org/licenses/LICENSE-2.0 + + Unless required by applicable law or agreed to in writing, software + distributed under the License is distributed on an "AS IS" BASIS, + WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + See the License for the specific language governing permissions and + limitations under the License. +--> + +# What is spatial index + +[A spatial index](https://gistbok.ucgis.org/topic-keywords/indexing) is a data structure that allows for accessing a spatial object efficiently. It is a common technique used by spatial databases. Without indexing, any search for a feature would require a "sequential scan" of every record in the database, resulting in much longer processing time. In a spatial index construction process, the minimum bounding rectangle serves as an object approximation. Various types of spatial indices across commercial and open-source databases yield measurable performance differences. Spatial indexing techniques are playing a central role in time-critical applications and the manipulation of spatial big data. + + + +# How does carbondata implement spatial index + +There are many opensource implementations for spatial indexing and to process spatial queries. CarbonData implements a different way of spatial index. Its core idea is to use the raster data. Raster is made up of matrix of cells organized into rows and columns(called a grid). Each cell represents a coordinate. And the index for that coodrinate is generated using longitude and latitude, like the [Z order curve](https://en.wikipedia.org/wiki/Z-order_curve). + +CarbonData rasterize the user data during data load into segments. A set of latitude and longitude represents a grid range. The size of the grid can be configured. Hence, the coordinates loaded are often discrete and not continuous. + +Below figure shows the relationship between the grid and the points residing in it. Black point represents the center point of the grid, and the red points are the coordinates at the arbitrary positions inside the grid. The red points can be replaced by the center point of the grid to indicate that the points lies within the grid. During data load, CarbonData generates an Index for coordinate according to row and column of the grid(in the raster) where that coordinate lies. These Indexes are the same as Z order. For the detailed conversion algorithm, please refer to the design documents of spatial index. + + + +When querying, the user enters the true space polygon coordinates, carbondata use the polygon and spatial region information passed in when creating a table build a quad tree. The nodes in the quad tree are composed of hash ids generated by the row and column information projected in the polygon area. When the query polygon area is not disjoint from the grid center point, the grid is considered selected. In the following figure, user select a quadrilateral polygon, The grid with the center point in the region will generate a quadtree. A list of line with continuous properties will be generated in the query process, like [97->97 99->99 102->102 104->111 120->120 122->123 151->151 157->158 159->159 192->208 210->210 216->216 225->225 228->229], each part of the list represents a continuous grid area. Carbondata use that line list to prune and filtered. About the detail can be search under https://issues.apache.org/jira/browse/CARBONDATA-3548 + + + + + +# Installation and Deployment Review comment: Modified ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [hidden email] With regards, Apache Git Services |
«
Return to Apache CarbonData JIRA issues
|
1 view|%1 views
| Free forum by Nabble | Edit this page |