Opened 5 years ago
Closed 5 years ago
#4436 closed defect (fixed)
ST_EstimatedExtent tests and PgSQL 12
Reported by: | pramsey | Owned by: | pramsey |
---|---|---|---|
Priority: | high | Milestone: | PostGIS 2.5.3 |
Component: | postgis | Version: | 2.5.x -- EOL |
Keywords: | Cc: |
Description
The new Pg!2 CTE behaviour allowing the planner inside the optimization fence changes the results of some of our regression tests. For example, estimated_extent.sql, if called without a MATERIALIZE
keyword will run the estimated extent query four times for this test (resulting in four warnings in the output) instead of the expected one.
with e as materialized ( select ST_EstimatedExtent('c1','g') as e ) select '#3391.1', round(st_xmin(e.e)::numeric, 2), round(st_xmax(e.e)::numeric, 2), round(st_ymin(e.e)::numeric, 2), round(st_ymax(e.e)::numeric, 2) from e; psql: WARNING: stats for "c1.g" do not exist ?column? | round | round | round | round ----------+-------+-------+-------+------- #3391.1 | | | | (1 row)
Note:
See TracTickets
for help on using tickets.
Fixed in 1dc840d0ac1c3379793fde6fbc9f816a79611158 and f4477cea937153e8bf2c55038025ddcf3d2258fe