Import error with long and similar filenames

Context

https://github.com/CartoDB/support/issues/967

Steps to Reproduce

1. Create a file named carto_long_filename_that_almost_matches_another_one_63chars_aaa.csv and import it into CARTO
2. Create a file named carto_long_filename_that_almost_matches_another_one_63chars_aab.csv and import it into CARTO

Current Result

Error 9999, the import log shows

2017-09-05 14:21:41 UTC: Exception: PG::Error: ERROR:  type "_carto_long_filename_that_almost_matches_another_one_63chars_aa" already exists in schema "xxxxx"

cc @juanignaciosl @rochoa

[javitonino]

Not exactly. Let me explain a bit more in depth:

• If the filename is longer than 63 chars, we truncate to 63 chars.
• If the table name is taken, we add a _1, _2, etc. suffix

So we support filenames of arbitrary length, we just convert them to something we can use on import. There is a bug (this one), where a filename is longer than 63 chars, and the first 62 chars match another table. In this case, it is expected that we import the file with no issues (they have a different name after all), but we are giving an import error instead.

This is clearly a bug, we should never throw an error on import, no matter the filename.

[kevin-reilly]

If we are truncating the 63rd character, we support file names with 62 characters. Right?

– Kevin Reilly SVP Product CARTO https://carto.com/ | [e-mail] [e-mail] [phone]

On Tue, Sep 5, 2017 at 11:04 AM, Javier Torres [e-mail] wrote:

We support any amount of characters and automatically truncate them. This is a specific bug with this truncation process, where the filenames are exactly the same except for the 63rd character.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/CartoDB/cartodb/issues/12732#issuecomment-327204443, or mute the thread https://github.com/notifications/unsubscribe-auth/ACy_X6dIO4sB3yd1j64NnQ4XD4h84WnKks5sfWL3gaJpZM4PNB6j .