Text file interpretation into various datatypes should be improved

[curtislisle]

Text files should have better default interpretations if this is possible. It is suggested to interpret as TSV or CSV initially

[jeffbaumes]

By this do you mean the handling of text files that have no registered extension (or no extension)?

[curtislisle]

I was actually referring to the default inference types of xxxx.txt files. This is a placeholder to remember that Luke has had trouble with text files being misinterpreted when the desired behavior would have been to interpret it as TSV.

I like the dialog that pops up asking for type clarification. Could we add TSV and CSV and option types, please?

On Aug 16, 2016, at 10:21 AM, Jeffrey Baumes notifications@github.com wrote:

By this do you mean the handling of text files that have no registered extension (or no extension)?

—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub #205 (comment), or mute the thread https://github.com/notifications/unsubscribe-auth/ACDZ9mnDvopKBpymtopaLJ1a8XBubXI5ks5qgfGkgaJpZM4Jlpb9.

[jeffbaumes]

I believe all delimited files are lumped under CSV and the separator (including tab) is auto-detected. This could just be a matter of adding txt to the list of extensions for CSV in Girder Worker. @danlamanna can you confirm and implement the change?