JSON not being copied on incremental sync

We’re trying to figure out estimates on timeline but need some more context on exactly when this issue began. Could you let me know the answers to these questions?

  1. Do you have normalization enabled? (Just want to triple check this so we don’t investigate in the wrong place)
  2. Was this table ever synced properly before reporting this issue? If so, do you know which Airbyte version and connector versions you were using, and did you make any changes to the Postgres schema around this time?

I have reproduced this issue

Airbyte v0.40.14
Source Postgres v1.0.16
Destination Snowflake v0.4.38

also:
Airbyte v0.40.10
Source Postgres 1.0.11
Destination Snowflake 0.4.38

Steps to reproduce:

  1. Inserted row with null column
  2. Updated row immediately, adding value to null column
  3. Sync to snowflake. Observe that the original insert and then the UPDATE transactions have same timestamp in raw table
  4. Observe UPDATE transaction does not appear in final table

Script I used for testing below and logs attached to this comment

import psycopg2from psycopg2.sql import NULLtry: connection = psycopg2.connect(user="postgres", password="postgres", host="127.0.0.1", port="5438", database="postgres") cursor = connection.cursor() postgres_insert_query = """ INSERT INTO test_table (id, code) VALUES ('row-test', None)""" cursor.execute(postgres_insert_query) postgres_update_query = """ UPDATE test_table SET code = 'update' WHERE id = 'row-test'""" cursor.execute(postgres_update_query) connection.commit() count = cursor.rowcount print(count, "Changes successfully added to table")except (Exception, psycopg2.Error) as error: print("Failed to add changes to table", error)finally: # closing database connection. if connection: cursor.close() connection.close() print("PostgreSQL connection is closed")