I'm running

meltano invoke dbt-postgres:run

with an apparently well configured transform on my PostgreSQL target using the

analytics

schema:

2022-06-21T16:50:25.279019Z [info     ] Environment 'dev' is active
16:50:28  Running with dbt=1.0.8
16:50:28  Found 1 model, 0 tests, 0 snapshots, 0 analyses, 165 macros, 0 operations, 0 seed files, 1 source, 0 exposures, 0 metrics
16:50:28  
16:50:29  Concurrency: 2 threads (target='dev')
16:50:29  
16:50:29  1 of 1 START view model analytics.events_per_day................................ [RUN]
16:50:29  1 of 1 OK created view model analytics.events_per_day........................... [CREATE VIEW in 0.28s]
16:50:29  
16:50:29  Finished running 1 view model in 1.01s.
16:50:29  
16:50:29  Completed successfully
16:50:29  
16:50:29  Done. PASS=1 WARN=0 ERROR=0 SKIP=0 TOTAL=1

and here is my model:

$ cat transform/models/tap_postgres/events_per_day.sql
SELECT
  CAST("timestamp" AS date) AS "timestamp"
  , event
  , distinct_id
  , COUNT(*) as "count"
FROM {{ source('tap_postgres', 'telemetry_events') }}
GROUP BY
  CAST("timestamp" AS date)
  , event
  , distinct_id

however, I don't see any new tables in the target. I would have expected this query to run for much longer by the way, since the source table is quite big. any way to debug what's happening here? (edit: solved ✔️ )

16:50:29  1 of 1 OK created view model analytics.events_per_day........................... [CREATE VIEW in 0.28s]

Created a view so no processing needed, that's why it's fast. Lots of debug options with DBT, check out

./transform/logs

if you'd like to really dig You have an analytics schema in postgres

You have an analytics schema in postgres

yep, that one is present, but apparently empty. I'll check the logs

try

select * from analytics.events_per_day

oh well, I think I might be missing some key PSQL knowledge here. indeed

select * from analytics.events_per_day

returns stuff, but I was confused because there were no tables listed in

SELECT * FROM <http://pg_catalog.pg|pg_catalog.pg>_tables

with that name, nor

select * from pg_matviews

. I was missing "views" (which are not "materialized views"), hence

select * from information_schema.views

). glad I sorted this out quickly, thanks!

np 😄

niceeeeeeeeee I understand everything now 🤯

@juan_luis_cano_rodriguez oh, so you are into Meltano too! Let’s talk about it on our next meeting of pydata! 🙂

oh well, I think I might be missing some key PSQL knowledge here

Yes, default relation type in DBT is a view, not a table. You can quickly list your views in

psql

with the

\dv

command. (e.g.

\dv analytics.*

@christoph awesome 😄 I rely on dbeaver but knowing

\dv

is nice 😄

I try not to leave the terminal if possible ... 😁 (Always trying to avoid the associated lag of switching from CLI to GUI ...)

yay @cesar_garcia_saez, the world is small! 😄