Has anyone experienced this before — you using `ta...
# troubleshooting
a
Has anyone experienced this before — you using
target-snowflake
(transferswise), you end up with a table that has a versioned column in it. In other words, the target table had a column
VALUE
, now it has both
VALUE
which contains values from before the previous run, and a column
VALUE_20230323_1125
? Is this a setting? I see something in the logs about making the column, but nothing about why this time it stuck, haha.
p
Checkout the advanced topics section on the hub page https://hub.meltano.com/loaders/target-snowflake#advanced-topics
Transfer wise has a pattern for handling schema changes. It sounds like what you’re seeing
a
ahhh. yeah, for some reason it’s seeing it as a version type change… thanks for the tip, gives me some thinking about how to force it not to do that.
Thanks!
FWIW, if anybody ever comes to this in the future: • For context, this happened in part because I switched from Stitch to Meltano, and started building on top of the old Stitch tables. • It looks like the Stitch loader treats type translations differently than the pipelinewise target-snowflake ◦ E.g., Stitch treats JSONB as
text
but target-snowflake wants
variant
◦ Stitch treats time-stamps without time zones as text, target-snowflake treats them as
time
type