Data mismatch in AWS DMS Bulk load vs CDC - postgresql

Question

I have a postgis database(source) for which i have done DMS and moved it to S3 bucket(target) in parquet file.
There is a column name point of datatype-geometry(point, 4326) in the source which converts to string after dms and looks like this in target-
-In bulk load it is like -point": "<Point srsName=\"EPSG:4326\"><coordinates>72.836903300000003,19.0823766</coordinates></Point>"
-In CDC it is like -point": "0101000020E610000051D77F4262855EC09591C4DCFFB54240"
I am able to get the coordinates back from the string in full load by string parsing but during CDC it is in some random hexadecimal string from which I am not aware how to decode back to the coordinates.

Answer 1

Welcome to SO.
In PostgreSQL in order to display geometries in a format other than WKB you have to explicitly state it in your query. In your case ST_AsGML:
SELECT ST_AsGML('0101000020E610000051D77F4262855EC09591C4DCFFB54240');
st_asgml
-------------------------------------------------------------------------------------------------------
<gml:Point srsName="EPSG:4326"><gml:coordinates>-122.0841223,37.4218708</gml:coordinates></gml:Point>
(1 row)
Or something like this if you wish to omit the namespace (as your example suggests):
SELECT ST_AsGML('0101000020E610000051D77F4262855EC09591C4DCFFB54240', 15, 0, '', '');
st_asgml
---------------------------------------------------------------------------------------
<Point srsName="EPSG:4326"><coordinates>-122.0841223,37.4218708</coordinates></Point>
(1 row)
See also: Converting geometries in PostGIS