Regex split file between substring

Question

I have a CSV file with data from multiple variables, and I would like to separate them. My file looks like this:

VARIABLE: GRP1.SGRP1.100:VAR1

Timestamp (LOCAL_TIME),Value
2018-07-18 13:52:09.100,25582
2018-07-18 13:52:49.900,24048
VARIABLE: GRP1.SGRP2.102:VAR1

Timestamp (LOCAL_TIME),Value
2018-07-18 13:52:09.100,25582
2018-07-18 13:52:49.900,24048

And I would like to split it on every occurrence of the substring "VARIABLE", producing two strings:

VARIABLE: GRP1.SGRP1.100:VAR1

Timestamp (LOCAL_TIME),Value
2018-07-18 13:52:09.100,25582
2018-07-18 13:52:49.900,24048

and

VARIABLE: GRP1.SGRP2.102:VAR1

Timestamp (LOCAL_TIME),Value
2018-07-18 13:52:09.100,25582
2018-07-18 13:52:49.900,24048

Something similar to VARIABLE[^V]+ would seem to work, but it should somehow terminate on the next occurrence, which I cannot figure out how. Thanks

@WiktorStribiżew Didn't work, placed everything on a single string — Bobisnotyouruncle, Aug 20 '18 at 08:25
I don't know Matlab, but you probably need to use the single line flag in order for this to work. — 41686d6564 stands w. Palestine, Aug 20 '18 at 08:27
Then try `matchStr = regexp(str,'VARIABLE[^V]*(?:V(?!ARIABLE)[^V]*)*','match')` — Wiktor Stribiżew, Aug 20 '18 at 08:27
@AhmedAbdelhameed `.` in [Matlab regex](https://www.mathworks.com/help/matlab/ref/regexp.html) *matches every character, including the newline*. There is no such a thing like *single line flag* there. — Wiktor Stribiżew, Aug 20 '18 at 08:42
@WiktorStribiżew Perhaps it didn't work because of the _default_ `noemptymatch` option then? — 41686d6564 stands w. Palestine, Aug 20 '18 at 08:48

Wiktor Stribiżew · Accepted Answer · 2018-08-20T08:55:25.113

You may use two approaches, matching and splitting.

Splitting is an easier approach, since the pattern will look like (?!^)(?=VARIABLE), but there is one caveat: Matlab regex expects non-empty matches by default (noemptymatch option is default). You need to pass the emptymatch option to regexp function for it to work:

splitStr = regexp(str,'\s*(?!^)(?=VARIABLE)','split', 'emptymatch')

Output:

splitStr = 
{
  [1,1] = VARIABLE: GRP1.SGRP1.100:VAR1

Timestamp (LOCAL_TIME),Value
2018-07-18 13:52:09.100,25582
2018-07-18 13:52:49.900,24048

  [1,2] = VARIABLE: GRP1.SGRP2.102:VAR1


Timestamp (LOCAL_TIME),Value
2018-07-18 13:52:09.100,25582
2018-07-18 13:52:49.900,24048
}

The (?!^)(?=VARIABLE) pattern matches any location in string that is not at the start of the string but that is immediately followed with a VARIABLE substring.

Alternatively, you may match VARIABLE and then any amount of non-Vs or Vs that are not followed with ARIBALE:

matchStr = regexp(str,'VARIABLE[^V]*(?:V(?!ARIABLE)[^V]*)*','match')

See the regex demo.

Details

VARIABLE - a VARIABLE substring
[^V]* - a negated character class matching 0 or more chars other thatn V
(?:V(?!ARIABLE)[^V]*)* - zero or more consecutive occurrences of
- V - a V char that is
- (?!ARIABLE) - ... not followed with ARIABLE
- [^V]* - 0 or more chars other than V.

Note it is "lexically" the same as VARIABLE(?:(?!VARIABLE).)* (with a tempered greedy token) or VARIABLE.*?(?=VARIABLE|$) (with lazy dot pattern and a mere positive lookahead), but is more efficient since it follows the unroll the loop principle. (Note that . in Matlab regex matches any char including newlines, so no need to use any additional flags when using these two patterns in Matlab).

Regex split file between substring

1 Answers1