Transaction text filter

[alekslyse]

Is it any function to run a regexp or something on the transaction name (payee etc) and description. I use curve in front of my apple pay and it adds a lot of text like crv and numbers that makes it very hard to read on YNAB. It would be great if we could clean up the strings before sending to YNAB.

[hpernu]

I am a Curve user as well and indeed we get all this stuff on front of Payee. However, this also happens for various other reasons (different payment processors add their own as well) and not all of them are handled by ynabber. In fact, in my case, most are not. So this is not really a problem in ynabber only but a general issue with payees.

I'd suggest you use payee management in YNAB itself instead of trying to fix this in ynabber. There is an option to combine similar payees to a single one and I think they added/are adding some AI based recognition to them. This may be a lot of work still but the only way to handle payees no matter where they are coming from.

[martinohansen]

@alekslyse we do have PayeeStrip which can be used to remove words from the Payee, I'm not sure if that will help in your case. Otherwise, what hpernu suggests is a good approach.

Can you share examples of the Payee and how you want it to look? We could make something like PayeeStrip, which accepts a regex command instead.

[hpernu]

I think I know since I am getting this same stuff from CSV imports from a couple of cards which are not handled by PSD2 compliant banks. In those I am getting CRV* -prepended to the actual real payee. But there are other payment processors as well which sometimes get prepended too. I've seen Vfi*, for example. Worst case is that I have two of them.

Even all PSD2 compliant banks might not actually export credit card transactions. I think I did not see those with Nordea. Even if it did, there I'd still recommend doing this mapping in YNAB instead. I think YNAB is getting better at recognizing payees and us trying to outsmart it might get in the way. I'd pass payees to YNAB for it to handle.

Even if we did strip the prefixes, due to limitations of most imports, the payee might be cut from the end. Since the prefix(es) take some characters in the beginning, longer payees are often shortened. This is probably a limitation of the banks in question which limit the number of characters in the payee information. We really cannot do anything about these -> better to do mapping in YNAB.

Curve has other interesting things as well like its GBiT-feature which I am constantly using. This will cause the transactions to reappear later on (and a refund in the original card). But handling them falls outside the scope of ynabber. I do have a solution for them myself though but is more a process rather than technology.

[hpernu]

If we are not going to implement regexp replacement for payee and/or memo, should we just close this issue with Won't Fix status?