4 ms·

Gotta wonder why the format isn't just a column separator char, a row separator char, and then all the data guaranteed not to have those two chars. Then you cou

Gotta wonder why the format isn't just a column separator char, a row separator char, and then all the data guaranteed not to have those two chars. Then you could save the thing by finding any two chars that aren't used in the data. I guess this is why we have a zillion formats.

m_eiman 5y ago
Then you can't edit or view it in a normal text editor, which is part of the appeal of CSV.
- ant6n 5y ago
  Maybe normal text editors should support the separators.
- alpaca128 5y ago
  That and the fact that the comma (or semicolon, as I prefer it due to the rare usage) is on every keyboard. A text editor can be made to handle those separators, but editing or processing the file can be more cumbersome than csv and tsv.
atoav 5y ago
I guess if you use an strange character nobody has on their keyboard, you could also just make it a binary format for efficiency
- Brometheus 5y ago
  See my sibling comment.
Brometheus 5y ago
ASCII Code 29: Group Separator ASCII Code 30: Record Separator ASCII Code 31: Unit Separator https://theasciicode.com.ar/ascii-control-characters/record-separator-ascii-code-30.html https://theasciicode.com.ar/ascii-control-characters/record-...
- srik 5y ago
  Thank you for pointing them out. I've browsed, and searched through that ascii table countless times, usually looking for something in particular, yet never realized that there were such a thing as separators before.
- kbouck 5y ago
  this doesn't directly address the ease-of-editing concern, but if you have csv (or tsv) and want ascii (0x1f, 0x1e) or unicode(U+241F, U+241E) separated (or vice versa), 'miller' [0] is a command line tool that supports converting among these formats with these one-liners: mlr --icsv --oasv cat input.csv # csv to asv mlr --iasv --ocsv cat input.asv # asv to csv [0] https://miller.readthedocs.io/en/latest/ https://miller.readthedocs.io/en/latest/ [1] https://miller.readthedocs.io/en/latest/file-formats/#csvtsvasvusvetc https://miller.readthedocs.io/en/latest/file-formats/#csvtsv...
  - jph 5y ago
    You're right, Miller is great for this, and it turns out that Miller already offers Unicode Separated Values (USV).
- cestith 5y ago
  It makes more sense to me in these days of ubiquitous encoding of in-band data that it would be easier to edit if we just used tabs and newlines. If you need a tab in your data, the standard could support either \t (requiring \\t if you need a literal backslash followed by a literal 't') or %09 or &09; or \x09 or something.
  - account42 5y ago
    If you require escaping then you may as well stick with csv - better to have the character that needs to be escaped to be common enough that people will run into it during testing.
    - cestith 5y ago
chimprich 5y ago
> Gotta wonder why the format isn't just a column separator char, a row separator char That's what the article proposes.

m_eiman 5y ago

Then you can't edit or view it in a normal text editor, which is part of the appeal of CSV.

ant6n 5y ago
Maybe normal text editors should support the separators.
alpaca128 5y ago
That and the fact that the comma (or semicolon, as I prefer it due to the rare usage) is on every keyboard. A text editor can be made to handle those separators, but editing or processing the file can be more cumbersome than csv and tsv.

atoav 5y ago

I guess if you use an strange character nobody has on their keyboard, you could also just make it a binary format for efficiency

Brometheus 5y ago
See my sibling comment.

Brometheus 5y ago

ASCII Code 29: Group Separator ASCII Code 30: Record Separator ASCII Code 31: Unit Separator https://theasciicode.com.ar/ascii-control-characters/record-separator-ascii-code-30.html https://theasciicode.com.ar/ascii-control-characters/record-...

srik 5y ago
Thank you for pointing them out. I've browsed, and searched through that ascii table countless times, usually looking for something in particular, yet never realized that there were such a thing as separators before.
kbouck 5y ago
this doesn't directly address the ease-of-editing concern, but if you have csv (or tsv) and want ascii (0x1f, 0x1e) or unicode(U+241F, U+241E) separated (or vice versa), 'miller' [0] is a command line tool that supports converting among these formats with these one-liners: mlr --icsv --oasv cat input.csv # csv to asv mlr --iasv --ocsv cat input.asv # asv to csv [0] https://miller.readthedocs.io/en/latest/ https://miller.readthedocs.io/en/latest/ [1] https://miller.readthedocs.io/en/latest/file-formats/#csvtsvasvusvetc https://miller.readthedocs.io/en/latest/file-formats/#csvtsv...
- jph 5y ago
  You're right, Miller is great for this, and it turns out that Miller already offers Unicode Separated Values (USV).
cestith 5y ago
It makes more sense to me in these days of ubiquitous encoding of in-band data that it would be easier to edit if we just used tabs and newlines. If you need a tab in your data, the standard could support either \t (requiring \\t if you need a literal backslash followed by a literal 't') or %09 or &09; or \x09 or something.
- account42 5y ago
  If you require escaping then you may as well stick with csv - better to have the character that needs to be escaped to be common enough that people will run into it during testing.
  - cestith 5y ago

chimprich 5y ago

> Gotta wonder why the format isn't just a column separator char, a row separator char That's what the article proposes.