PurujitChaugule commented on a change in pull request #79:
URL: https://github.com/apache/carbondata-site/pull/79#discussion_r611622451
##########
File path: src/site/markdown/spatial-index-guide.md
##########
@@ -32,12 +32,28 @@ Below figure shows the relationship between the grid and
the points residing in
Carbon supports Polygon User Defined Function(UDF) as filter condition in the
query to return all the data points lying within it. Polygon UDF takes multiple
points(i.e., pair of longitude and latitude) separated by a comma. Longitude
and latitude in the pair are separated by a space. The first and last points in
the polygon must be same to form a closed loop. CarbonData builds a quad tree
using this polygon and spatial region information passed while creating a
table. The nodes in the quad tree are composed of indices generated by the row
and column information projected in the polygon area. When the grid center
point lies within the polygon area, the grid is considered as selected. In the
following figure, user selects a quadrilateral shaped polygon. The grid at the
center of the region is chosen to build a quad tree. Once tree is build, all
the leafs are scanned to get the list of range of indices(with each range
consisting of minimum index and maximum index in the range). All th
e indices starting from minimum to maximum in each range forms the result.
-The main reasons for faster query response are as follows :
-* Data is sorted based on the index values.
-* Polygon UDF filter is pushed down from engine to the carbon layer such that
CarbonData scans only matched blocklets avoiding full scan.
+There are some other UDFs supporting more filter conditions in the query,
including Polygon List, Polyline List, and spatial index range list.
+
+Polygon List UDF takes multiple polygons(i.e., a set of points) and operation
type for combining polygons. Only `OR` and `AND` are supported at present,
operation 'OR' means union of multiple polygons and 'AND' means intersection of
that, shown as the following figure. Then CarbonData gets the list of range of
indices from the combined region by quad tree, which is the same processing as
Polygon UDF.
+
+
+
+Polyline List UDF takes multiple polylines(i.e., a set of points) and buffer
in meter. CarbonData first converts polyline to polygon and then gets the list
of range of indices from these polygons. The processing is the same as Polygon
UDF and return all the data points lying within the buffer region of polylines.
+
Review comment:
can mention the unit of measurement i.e. meters, used for buffer size in
PolyLine UDF
--
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:
[email protected]
