Bug 127718

Summary:	Calc cannot split columns with multi-character delimiters
Product:	LibreOffice	Reporter:	david.cortes.rivera
Component:	Calc	Assignee:	Not Assigned <libreoffice-bugs>
Status:	RESOLVED NOTABUG
Severity:	normal	CC:	oliver.brinzing, vsfoote
Priority:	medium
Version:	Inherited From OOo
Hardware:	All
OS:	All
See Also:	https://bugs.documentfoundation.org/show_bug.cgi?id=132414
Whiteboard:
Crash report or crash signature:		Regression By:
Attachments:	split_multi_char_delim

Description david.cortes.rivera 2019-09-23 12:02:36 UTC

Steps to reproduce:
- Create some delimited file with multi-character delimiters, e.g.
col1||col2||col3
a||b||c
- Open said file with LO Calc.
- In the text import dialog, select "Separated by", check "Other" and write "||" in there as separator.
- Click OK and check the file inside Calc.

Expected behavior: should separate columns every time there's two bars (||).

Actual behavior: separates columns every *single* bar, thus creating empty columns (can be fixed by checking "Merge delimiters", even though it's only one delimiter).

Comment 1 david.cortes.rivera 2019-09-23 15:39:01 UTC

By the way, the bigger issues are that:
- It's not possible to split by non-repeated characters correctly, e.g.
a|,|b|,|c
text1|,|text2,text_after_comma|,|text3|text_after_bar
- Even if it's the same character repeated multiple times, it will still split by single character, e.g.:
a||b||c
text1||text2|text_after_bar||text3

Comment 2 Oliver Brinzing 2019-09-23 17:36:49 UTC

reproducible with:

Version: 6.4.0.0.alpha0+ (x64)
Build ID: 71ef762f21ada8c25aad2183065478171e985e8c
CPU threads: 4; OS: Windows 10.0; UI render: default; VCL: win; 
Locale: de-DE (de_DE); UI-Language: en-US
Calc: threaded

Comment 3 Oliver Brinzing 2019-09-23 17:39:48 UTC

Created attachment 154392 [details]
split_multi_char_delim

it's only possible to hide the unwanted columns

Comment 4 V Stuart Foote 2019-09-23 19:04:54 UTC

No, using 'Other' with a multiple character separators is then collapsed using the 'Merge delimiters' checkbox.

Caveat is that if you use or work with strings that actually contain the separator character alone it would parse each out as additional columns.  But any utf-8 Unicode codepoint can be used, not limited to ASCII so make a good choice.

And unfortunately there is no awk/nawk style "string" separator; the Merge delimiter does not act on the string--just each character of the string. 

But that would be an RFE

Comment 5 V Stuart Foote 2019-09-23 19:11:08 UTC

Oh, and if using a Unicode glyph for the delimiter, remember to adjust the 'Character set:' encoding to Unicode (UTF-8) in the Text Import dialog.