CSV: match group references get mixed up when multiple ifs match a record

[amoralesch]

When reading CSV and rules files and using REGEX expressions, if more than one if block matches the same record, the REGEX group index gets mixup.

Example:

• CSV file:

date,desc
2019-01-01,SUFFIX First
2019-02-01,SUFFIX Text 1 - Text 2

• Rules file:

skip          1

fields        date,desc
separator     ,
date-format   %Y-%m-%d
description   Test

if %desc SUFFIX (.*)
  account1      account:\1
  amount1       1
  account2      equity

if %desc SUFFIX (.*) - (.*)
  account1      account:\1:\2
  amount1       1
  account2      equity

Expected Result

2019-01-01 Test
    account:First               1
    equity

2019-02-01 Test
    account:Text 1:Text 2               1
    equity

Current Result

2019-01-01 Test
    account:First               1
    equity

2019-02-01 Test
    account:Text 1 - Text 2:Text 1               1
    equity

Note

If the rules file is modified with the following:

if %desc SUFFIX (.*) - (.*)
  account1      account:\2:\3
  amount1       1
  account2      equity

(note the index for matching groups (\2 and \3), then the result works as expected.

More info

HLedger version: hledger 1.32.2, mac-aarch64

[Background: wish #2009, pr #2087.
Workaround: don't allow multiple ifs with capture groups to match the same CSV record. So if you use capture groups, make those regexps specific enough that at most one of them will match any given CSV record.]

[amoralesch]

Related? #2009

[amoralesch]

FYI: @jmtd

[jmtd]

Interesting: I'll take a look, thank you.

[simonmichael]

I'd like to get this one in today's release, and have time to take a look if needed.

[simonmichael]

I left some comments on the original PR #2087.

the workaround for now is when you use capture groups, make those regexps specific enough that at most one of them will match any given CSV record.

[jmtd]

I'd like to get this one in today's release, and have time to take a look if needed.

Sorry -- I've just seen your comment -- but I'm not able to take a look at the moment, I've got too much work on. I might have a chance to start looking at the end of this weekend (FOSDEM/travel back from FOSDEM), but I don't know how long it will take to fully address.