csvsql: "no such table" error only thrown when force-quitting

I am querying a data file:

csvsql --query 'select max(video_session_started_at) from data' tmp_correct_ids.csv

I did this, naively, by adapting the existing code, assuming that data would be a generic name for the table that’d be generated in memory after reading from the CSV.

I was wrong. It seems the table name is inferred from the CSV filename without suffix. I later learned that from carefully reading the manpage.

However, earlier, I only found out after I killed the command:

^C(sqlite3.OperationalError) no such table: data
[SQL: select max(video_session_started_at) from data]
(Background on this error at: http://sqlalche.me/e/e3q8)

Would it be possible to throw this error and stop csvsql once it occurs? The command just seemed to hang for me until I quit it.

Using csvsql 1.0.4 under Ubuntu 18.04 with Python 3.6.

[dannysepler]

just an update here, i’m going to bow out of this one. Sorry! It doesn’t seem like a bug so much as a speedup feature. While that would still be fun, the only ways I can rationalize doing it are pretty messy. (Like you said, the table name inference is easy. But how do we know what tables are being are being queried for? Seems like we’ll need a SQL parser of some sort.)

To me an interesting question as well is around what happens when you Ctrl-C a big table read, and it gets caught in the “StopIteration” here. Somehow, it is still able to successfully continue the command!

Just leaving some food for thought for whoever picks this up next.

[jpmckinney]

Thanks in advance, @dannysepler!