I all, I have been wrestling with this issue for a few days now and I cant seem to find the solution. When I use tap-csv or tap-s3-csv it doesnt seem to be taking the column names of the csv with it. I have tried loading with target-redshift and target-postgres (screen shot is from postgres) and both return similar errors of 'cannot create a table without column names'. Attached is the config files for tap-s3-csv and target-postgres (what I have been using in my most recent attempts). I am still very new to meltano so I am sure it is probably an obvious mistake, but no matter what I try I am continuously getting Column must be constructed with a non-blank name or assign a non-blank .name before adding to a Table. Attached is also the csv I am trying to upload. Just a 'dummy test' csv. Please Help

Hey Matt! What do you see if you run

meltano invoke tap-s3-csv

?

let me know if you need any additional logs from the command. Thank you for the help!

Is there a

{"type": "SCHEMA", ...}

message above those lines?

yes, here it is

Aha so there's a column there with

""

. Can you try looking at the output of

meltano select tap-s3-csv --list --all

?

oh I see, but my csv doesnt have an empty col? is it a formatting issue with the csv? Also thank you so much for the help thus far

Yeah, I'm not sure how that empty col is sneaking in the schema. Let's try using select:

extractors:
- name: tap-s3-csv
  select:
  - names_table.Age
  - names_table.Name

that is very handy! I will be sure to use select. I updated it and it seems I am onto a new problem. it looks like it is inserting into a random string of characters instead of names_table. Where is it getting this from?

Of course! I think that's only the auto-generated temporary table name. Is is not creating a

names_table

by the end of the run?

I dont think so, the run ends shortly after that message

It seems to have completed successfully so you can probably confirm by running

select * from names_table

against your postgres db.

doesnt seem to have made the table, it is possible I am missing something though

should I be making the table before

nope, the loader handles that automatically for you Hmm, maybe there's something weird going on with state, so can you try using the

--full-refresh

flag of the

run

command?