Skip to content

New table does not always load (all) data #37

@skjbulcher

Description

@skjbulcher

Hi,

I'm playing around with singer.io for the first time. I've set up a tap-mysql to target-bigquery script. My test database contains two rows that I'm using as a playground.

I noticed that target-bigquery is capable of creating the BigQuery tables based on the input from tap-mysql, which I love 😍 I wasn't looking forward to creating tables manually. However, when I examined the contents of a new table after running target-bigquery, it consistently lacks one or both records from the tap. The target, however, always reports that it "Loaded 2 row(s) into testganger:template /projects/ebs-it/datasets/testganger/tables/template", regardless of the contents of the table.

Curious, I ran the target several times on an existing database, and in every case two records were always created. The issue appears to be limited to the creation of new tables.

Steps to reproduce

  1. Set up tap-mysql with a small table - two rows.
  2. Run tap-mysql and output the state to a file (let's call it state.log)
  3. Run target-bigquery -c config.json < state.log to load the data into BigQuery
  4. Run target-bigquery -c config.json < state.log to load the data into BigQuery a second time.

What happened
The table is created properly in BigQuery. Runing select * from <table> results in the display of between 2-4 records.

What I expected
Runing select * from <table> results in the displays exactly 4 records every time.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions