Support for tsv file ingest for tabular data

[posixeleni]

We currently do not support special ingest for tabular files that are in a tab-delimited format (.tsv). Spoke w/ @mcrosas that this is something we should also support if we already can handle .csv format which similarly looks at the first row of the table to grab variable metadata.

[philippconzett]

As for tabular data, in addition to the original file format we recommend our users to provide a tabulator separated Unicode utf-8 .txt file. Those seem not be be ingested properly in Dataverse. Here are to examples that can be tested:

http://dx.doi.org/10.18710/SJ4OQE
http://dx.doi.org/10.18710/4MCRQL

Best, Philipp

[pdurbin]

@philippconzett thanks. To be clear you mean that "Spring_migration_arrival_dates_Acrocephalus_schoenobaenus.txt" (MD5: fcc151933f34034bfd1ea6b2ff3d3186), for example, is a tab separated values file from https://opendata.uit.no/dataset.xhtml?persistentId=doi:10.18710/4MCRQL that you would like to pass ingest. (At https://opendata.uit.no/dataset.xhtml?persistentId=doi:10.18710/SJ4OQE I don't actually see any tab separated .txt files.)

[philippconzett]

Sorry about that. I should have linked to a tab separated uft-8 .txt file that has not been properly ingested. Here is an example:
http://hdl.handle.net/10037.1/10296
See e.g. the file: Database fiskesammensetninger for publisert versjon Maal og Minne 2016_1.txt

[pdurbin]

@philippconzett ah, yes, I see what you mean. Thanks.

[pdurbin]

As I mentioned at http://irclog.iq.harvard.edu/dataverse/2016-11-14#i_44974 a workaround for me has been to convert my TSV file to a CSV file by using https://github.com/brendano/tsvutils . I cloned the repo and ran cat /tmp/irclog.tsv | python tsv2csv > /tmp/irclog.csv

[pdurbin]

@oscardssmith does the work you're doing in #3930 help us get support for ingest of TSV files?