javascript regex with [x|y|z|xy|yz]

Question

I have a regex:

\[[^\[\]]*\]\s*[<>|<=|>=|=|>|<]\s*'?"?\w*'?"?

this basically just parse a equation like:

[household_roster_relationships_topersona_nameadditionalpersono] = "1"

it works good, with

'=','>','<'.

but when the equation has

'<=','>=','<>'.

the parse stop at first character of

'<=','>=','<>'.

I have created a demo on regex101

How can I correct regex so it will work in this situation?

You should learn about character class: `[<>|<=|>=|=|>|<]` == `[=><|]` — Thomas Ayoub, Feb 16 '17 at 14:52

score 2 · Answer 1 · edited Sep 24 '17 at 18:54

2

Just change your char class for an alternation:

\[[^\[\]]*\]\s*(<>|<=|>=|=|>|<)\s*'?"?\w*'?"?
               ^              ^

See demo

edited Sep 24 '17 at 18:54

Graham

7,431
18
59
84

answered Feb 16 '17 at 14:54

Thomas Ayoub

29,063
15
95
142

Probably make it a non-capturing group as well! – ibrahim mahrir Feb 16 '17 at 14:55
1

@ibrahimmahrir one can, but this will not change the match – Thomas Ayoub Feb 16 '17 at 14:56
It will enhance the performance though! OP may learn something! – ibrahim mahrir Feb 16 '17 at 14:57
See the section **Non-Capturing Groups** [**here**](http://www.regular-expressions.info/brackets.html). It says _you can optimize ..._! – ibrahim mahrir Feb 16 '17 at 15:02
1

@ibrahimmahrir it doesn't says if it's about speed opti or data volume opti, I'll run some tests – Thomas Ayoub Feb 16 '17 at 15:05
Just log the return of the match and you'll understand! – ibrahim mahrir Feb 16 '17 at 15:06
1

@ibrahimmahrir I *know* the difference in the returned data, I have no clue about performance difference – Thomas Ayoub Feb 16 '17 at 15:08
In this example the performance is pretty much the same because there is only one group of at most 2 characters. But it's not a good practice to use capturing groups if they won't be used. That's why I said OP could learn something (to use in future regular expressions)! – ibrahim mahrir Feb 16 '17 at 15:10

score 1 · Accepted Answer · answered Feb 16 '17 at 14:54

1

You need to replace the character class with a grouping construct.

Use

\[[^[\]]*]\s*(?:<>|[<>]=|[=><])\s*['"]?\w*['"]?
             ^^^              ^

See the regex demo. The (?:<>|[<>]=|[=><]) non-capturing group (only used for grouping subpatterns) matches either <>, <=, >=, =, > or <.

Note I reduced some alternative branches to make the pattern a bit more compact. Also, I think you just want to match either ' or " at the end, so, a mere ['"]? (1 or 0 ' or ") should be enough.

Also, you do not need to escape a [ inside a character class ([[] matches a single [) and you do not need to escape ] outside a character class, it matches a literal ] symbol.

answered Feb 16 '17 at 14:54

Wiktor Stribiżew

607,720
39
448
563

Hi, that works good. with the example in the demo. I have another example that break the regex. [diagnostic_procedures_performed] = 'UNDEFINED_CODE' and [other_procedure_date] > "01-01-1901" how would I fix too? – Peter Huang Feb 16 '17 at 16:55
Replace `\w*` with `[\w-]*` (0+ word or `-` symbols) or with `[^"']*` (0+ chars other than `'` and `"`). The latter may be more generic. – Wiktor Stribiżew Feb 16 '17 at 17:06

m87 · Answer 3 · 2017-02-16T15:03:16.313

0

You need to use () instead of [] for multiple word character class selection ...see updated regex

edited Feb 16 '17 at 15:03

answered Feb 16 '17 at 14:57

m87

4,445
3
16
31

javascript regex with [x|y|z|xy|yz]

3 Answers3