查找并用sed替换IP地址的最后一段

Question

I have bunch of json files that contains web events. Each event contains lots of stuff and I'm trying to do ip address anonymization (replacing last segment of ip addresses with 0) with sed .

In short:

How to find substrings like "ip":"34.542.3.34" from json files and transform them to "ip":"34.542.3.0" ?

Attempts:

1. Resetting starting point with \\K

sed -e 's/"ip":"[0-9]{1,3}\\.[0-9]{1,3}\\.[0-9]{1,3}\\.\\K[0-9]{1,3}/0/g' file.json

This would work but unfortunately sed doesn't seem support reseting starting point.

2. Negative lookbehind

sed -e 's/(?<="ip":"[0-9]{3}\\.[0-9]{3}\\.[0-9]{3}\\.)([0-9]{3})/0/g' file.json

This would also work but negative lookbehind doesn't seem to support non-fixed-width assortions. So [0-9]{1,3} is not supported and hence this won't work.

3. Matching groups

Third idea was to use matching groups and do something like this

sed -e 's/("ip":"[0-9]{1,3}\\.[0-9]{1,3}\\.[0-9]{1,3}\\.)|([0-9]{1,3})/\\1\\20/g' file.json

But couldn't figure out how this would actually work with sed .

4. Writing regex for every possible length option separately

This would probably work but it would make regex too long and hardly readable. I would like to find more convenient and clean solution.

Any suggestions?

Answer 1

使用GNU sed：

sed -r 's/("ip":"[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}\.)[0-9]{1,3}(")/\10\2/' file

Answer 2

如果没有-r -option：

sed -e 's/\("ip":"[[:digit:]]\{1,3\}\(\.[[:digit:]]\{1,3\}\)\{2\}\.\)\([[:digit:]]\{1,3\}\)"/\10"/g' tst.json