CODEWITHSUNDEEP
pythonregex
beruic
beruic

Reputation: 5866

Match numbers not preceded by string

I'm analysing some very large log files using Python regex. I need to substitute every number in the log file, except the numbers that are preceded by 'java:' (the log files are made by a java program).

This means that given we have a line saying:

This is a bogus test line with limit=300 doing 53 rounds and the error is in (Abc.java:417) and some more

The numbers 300 and 53 should be replaced, but not 417.

I filter on a line basis, and it should be noted that not all lines contain java:[number].

The closest I have gotten is ((?<!java:)[0-9]+)

Upvotes: 6

Views: 902

Answers (1)

Andrew Cheong
Andrew Cheong

Reputation: 30273

Probably what's happening with

((?<!java:)[0-9]+)

is that, sure, the match at this point,

java:
     ^

fails, but then at _this point,

java:4
      ^

succeeds, because indeed, ava:4 is not java:.

You'll just need to add one more negative lookbehind,

((?<!java:)(?<![0-9])[0-9]+)
           ^^^^^^^^^^

so that only "complete" numbers are considered.

Upvotes: 4

Related Questions