[FEATURE] Add tsv (tab separated) as save format

[keiranmraine]

Is your feature request related to a problem? Please describe.
TSV files are used extensively in bioinformatic spaces.

Describe the solution you'd like
Support SAVE direct to tsv.

Describe alternatives you've considered
Outputter following SAVE AND LOAD

[kvnkho]

Just adding notes, I was testing this and the issue is we don't allow filename with .tsv extension. We should support it. For a lot of the engines, it's jut changing the delimiter of the method to read csvs

[goodwanghan]

We can support TSV already

[keiranmraine]

FYI, this doesn't handle windows file paths:

File <string>:1
    "c:\Users\XXX\cleaned.tsv"
                                                                                                      ^
SyntaxError: (unicode error) 'unicodeescape' codec can't decode bytes in position 2-3: truncated \UXXXXXXXX escape

I had to do:

fe_output = f_output.replace("\\", "\\\\")

%%fsql
SELECT df_links.uuid, df_pcawg.*
    FROM df_pcawg
    INNER JOIN df_links ON df_pcawg.icgc_donor_id = df_links.DO
SAVE OVERWRITE CSV "{{fe_output}}" (sep="\t", header=TRUE)

[keiranmraine]

Changing the last directive to SAVE AND USE OVERWRITE CSV "{{fe_output}}" (sep="\t", header=TRUE) results in an error:

File c:\Users\XXX\.venv\lib\site-packages\fugue\workflow\workflow.py:1518, in FugueWorkflow.run(self, *args, **kwargs)
   1516         if ctb is None:  # pragma: no cover
   1517             raise
-> 1518         raise ex.with_traceback(ctb)
   1519     self._computed = True
   1520 return DataFrames(
   1521     {
...
    231         pdf = _safe_load_csv(
    232             p.uri, **{"index_col": False, "header": None, "names": columns, **kw}
    233         )

ValueError: columns must be set if without header

[goodwanghan]

FYI, this doesn't handle windows file paths:

File <string>:1
    "c:\Users\XXX\cleaned.tsv"
                                                                                                      ^
SyntaxError: (unicode error) 'unicodeescape' codec can't decode bytes in position 2-3: truncated \UXXXXXXXX escape

I had to do:

fe_output = f_output.replace("\\", "\\\\")

%%fsql
SELECT df_links.uuid, df_pcawg.*
    FROM df_pcawg
    INNER JOIN df_links ON df_pcawg.icgc_donor_id = df_links.DO
SAVE OVERWRITE CSV "{{fe_output}}" (sep="\t", header=TRUE)

This is actually a correct behavior. It may be tricky to understand, but let me use a native python code to illustrate:

The eval here tries to parse the expression with double quotes, which is very similar to fsql. And you see I can reproduce the error you had. Let me know if you can understand what I try to explain.

[goodwanghan]

@keiranmraine please see the issue 332 ^^^

[goodwanghan]

Closing