jonas_kalderstam
10/21/2022, 1:43 PMint4range into two new columns x_start , x_end, as part of the loading. Has anyone solved this with existing taps?jonas_kalderstam
10/21/2022, 1:58 PMSELECT x FROM table
to do this
SELECT lower(x) as x_start, upper(x) as x_end FROM tablevisch
10/21/2022, 2:30 PMname: tap_postgres:
config:
additional_queries:
- name: xstart_xend
sql: SELECT lower(x) as x_start, upper(x) as x_end FROM table
I don't think that's really what you're after here as you're looking to apply this to multiple queries right?
If you're trying to use lower(x) and upper for incremental loads or something then it's a different question 🤷jonas_kalderstam
10/21/2022, 2:31 PMjonas_kalderstam
10/21/2022, 2:31 PMjonas_kalderstam
10/21/2022, 2:31 PMst_text(...) to be runvisch
10/21/2022, 2:32 PMjonas_kalderstam
10/21/2022, 2:33 PMvisch
10/21/2022, 2:34 PMjonas_kalderstam
10/21/2022, 2:34 PMvisch
10/21/2022, 2:35 PMvisch
10/21/2022, 2:35 PMjonas_kalderstam
10/21/2022, 2:36 PMjonas_kalderstam
10/21/2022, 2:36 PMaaronsteers
10/21/2022, 4:33 PMaaronsteers
10/21/2022, 4:36 PMjonas_kalderstam
11/04/2022, 10:02 AMjonas_kalderstam
11/04/2022, 10:03 AMaaronsteers
11/04/2022, 7:36 PMaaronsteers
11/04/2022, 7:39 PMvisch
11/04/2022, 7:42 PMaaronsteers
11/04/2022, 7:50 PMCREATE (temp?) TABLE temp_foo AS SELECT ... LIMIT 0 and scanned metadata from that. Neither of these solutions is very elegant.
That said, I just realized now that (at least in theory), this cost could be paid during discovery and not during stream time if the schema is cached into the catalog.json and if the sync-mode operation leverages it when available. 🤔 In theory, 'discovery' could also be run with additional permissions - and additional patience for the added scan/compile costs.