Closed KoalaGeo closed 4 months ago
If you have SDO_GEOMETRY
columns in the table, I would expect bad performance. Reading such tables requires a round trip between PostgreSQL and Oracle for each table row. This limitation has been lifted for LOB columns in version 2.6.0, but it still applies to geometry data.
The only way to improve the performance is to reduce the network latency between Oracle and PostgreSQL.
Thank you for that clarification 👍
I'll have a go making the same tables without SDO_GEOMETRY
but including a lat/long, then making the geometry column in PostGres.
Right; a view might be a good workaround.
Hi @laurenz,
Hoped you might be able to offer some advice. I'm running Oracle & Postgres and am trying to populate postgres with ~40 geospatial tables each with 1 - 2M rows of data.
IMPORT FOREIGN SCHEMA "PUBLISHED"
CREATE MATERIALIZED VIEW
for the tables I'm interested inWITH NO DATA
REFRESH MATERIALIZED VIEW
This is done via flyway & GitLab CI, with the refresh function called at the end (and on schedules) to populate the views.
This refresh job is currently taking nearly 3 hours hours to run. Do you know if there's as way I could speed this up?
Many thanks