Elasticsearch regexp not marching but the regexp pattern is correct

Question

"regexp": {
  "commonStructure": {
    "value": "[^>]*?[,<]PRP[,>][^\\<]*?\\s[^>]*?[,<]VERB[,>].*",
    "flags": "ALL"
  }
}

The pattern is correct and its working in js and python

And her is the example string

<PRP,PRON>{I} <VBD,VERB>{worked} <ADVPL,IN,ADP>{in} <NNP,PROPN>{London} <IN,ADP>{at} <ADVTMRI,RB,ADV>{first} <PUNCT>{,} <CC,CCONJ>{but} <PR,PRON>{it} <VBD,VERB>{was} <RB,ADV>{not} <JJ,ADJ>{easy} <TO,PART>{to} <VB,VERB>{make} <NN,NOUN>{money} <ADVPL,RB,ADV>{there} <PUNCT>{.}

In this pattern, I want to get all records that have first [PRON] and the next neighbor is the [VERB]

note that in this pattern with first [PRON] i can pass also the word in this case {I}

If the regex works in Python and JS, it does not mean it will work in other regex flavors. \s is not supported by the Lucene regex engine. Try "[^>]*[,<]PRP[,>][^\\<]*[ \t\r\n][^>]*[,<]VERB[,>].*". BTW, what did you match with [^\\<]? — Wiktor Stribiżew
– Wiktor Stribiżew, Commented Aug 16, 2018 at 11:42
Here is the supported regexp syntax for Lucene: elastic.co/guide/en/elasticsearch/reference/current/… — Val
– Val, Commented Aug 16, 2018 at 11:46
Tanks @WiktorStribiżew jan your suggestion is really helpful and its fixed my issue, But it's so strange that \s isn't working in regex — ashmna
– ashmna, Commented Aug 16, 2018 at 11:49
@WiktorStribiżew [^\\<] it is a typo from copy-pasting it must be [^\<] and yes I know that the correct one is [^<] — ashmna
– ashmna, Commented Aug 16, 2018 at 11:53
@WiktorStribiżew I Tink this sone about you youtube.com/watch?v=IB7s3YlbU9U — ashmna
– ashmna, Commented Aug 16, 2018 at 11:55

Wiktor Stribiżew · Accepted Answer · 2018-08-16 11:56:16Z

1

Lucene regex engine does not support common NFA regex shorthand character classes.

To match a whitespace, you may use your own bracket expression like [ \t\r\n], or [ \f\n\r\t\v].

To match any char but <, use [^<], do not escape < as it is not a special regex char.

So, you may use

"[^>]*[,<]PRP[,>][^<]*[ \t\r\n][^>]*[,<]VERB[,>].*"

answered Aug 16, 2018 at 11:56

Wiktor Stribiżew

631k41 gold badges502 silver badges633 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

Elasticsearch regexp not marching but the regexp pattern is correct

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related