Android Studio Kotlin regex different than expected

Question

I'm having a problem with a specific regex that is returning a different value than expected when running in Android Studio.

Scenario:

The code is simple:

val regex = "(?<=N|E|\\G)\\d{2}(?=\\d*$)".toRegex()
print("${regex.findAll("N2032354345").count()}")

This should print 5 as there are 5 matches in this string (https://regex101.com/r/6PDbkI/1) and if we run in on Ideone.com or in Kotlin Playground, the result is the expected 5.

However, in Android Studio, the result is 1:

Theory:

It seems that the regex in Android Studio is failing to use the \G operator (which might be related to Kotlin split with regex work not as expected)

Anyone faced the same problem? Is there any way to change the regex to a similar one that isn't failing in Android Studio? Am I missing some setting?

Sounds more like a bug to me, considering it works fine in the playground — Zoe - Save the data dump
– Zoe - Save the data dump, Commented Nov 9, 2018 at 14:27
That's what I thought @Zoe. Do you happen to know what is the best place to report such issue? — Pedro Oliveira
– Pedro Oliveira, Commented Nov 9, 2018 at 14:28
And that's correct @WiktorStribiżew. But I'm developing for Android Studio, which is not returning 5, but 1 instead (which is NOK) — Pedro Oliveira
– Pedro Oliveira, Commented Nov 9, 2018 at 14:30
I think you may use a workaround like "(?<=[NE]\\d{0,100})\\d{2}(?=\\d*$)" (you may adjust the 100 value). — Wiktor Stribiżew
– Wiktor Stribiżew, Commented Nov 9, 2018 at 14:31
I can confirm that that workarround works @WiktorStribiżew. But it seems to be an ugly one :/ — Pedro Oliveira
– Pedro Oliveira, Commented Nov 9, 2018 at 14:32

Wiktor Stribiżew · Accepted Answer · 2018-11-09 14:44:38Z

5

Android Pattern documentation lists \G as a supported operator:

\G The end of the previous match

Hence, it sounds like an Android Studio bug.

Until it is fixed, you may use a work around for your scenario that involves just a dozen digits in the input:

val regex = "(?<=[NE]\\d{0,100})\\d{2}(?=\\d*$)".toRegex()

The pattern matches:

(?<=[NE]\d{0,100}) - a position that is immediately preceded with N or E and 0 to 100 digits
\d{2} - two digits
(?=\d*$) - that are followed with 0 or more digits to the end of the string.

answered Nov 9, 2018 at 14:44

Wiktor Stribiżew

631k41 gold badges502 silver badges632 bronze badges

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

1 Comment

Thx @WiktorStribiżew +1 for saving my day