Closed alexeygrigorev closed 7 years ago
(although there's an issue with parsing the last column)
Thank you for reporting your issue. Indeed, #31 solves the issue, but we are waiting for the PR issuer to remerge so we can run the tests on the PR before merging into master
.
Most of the regression tests assume all data is double-quoted because this is what I do for most of the data files I used in a production environment. paratext
supports backslash-escape sequences so in theory any arbitrary byte sequence can be represented.
If you have a very messy CSV file, you can use: paratext.serial.write_frame
, which will write the data out using a configurable backslash escaping scheme (writing arbitrary 8-bit, printable ASCII, UTF-8, etc). In fact, the regression tests generate arbitrary UTF-8 and byte data, save in all possible formats, and read it back in. However, the key assumption to get this to work is that all non-numeric data is backslash-escaped.
This issue has been resolved in the latest master.
I'm reading the following csv file:
But paratext reads it as following:
The uuid conversion is totally unexpected - and the issue persists even if I say
text_names=['uuid']