No, the database is incidental. Let's do a remedial debugging review to see why.

You notice that '0' is getting printed as ''. Having seen a zero come back when you ran the query in a command-line tool, you suspect that the issue isn't the database, so you work from the end backwards. You note that the path between the '0' that you have and the print that is printing as an empty string includes a join, a map, and a call to trim.

Knowing (or strongly suspecting) that join and map don't zap zeros, you direct your attention to trim. What comes back when we pass a '0' to trim()? An empty string!

Voila. The database (and most the rest of the script) is now out of the loop completely. You have isolated the problem to a single subroutine.

Assuming that you're unable to discern how trim() could possible turn a zero into an empty string, you nonetheless have a small snippet to post.