Is there a way to pull a statement date into renaming the document title?

[BlankSlate56]

I want to use paperless to scan in statements. I was hoping it would be able to recognize the statement date and then be able to use that date when renaming the file. for example, if I have a US Bank Statement from 2/1/2021, I want the name of the file to read "USB Statement 2-1-21.pdf". I don't want to use the "added" or "created" date because those dates don't match the statement date. Is this possible?

Also, I want to rename the files because I'll be pulling files into my bookkeeping software, and so it'll be much easier to find files by filename than going into paperless, finding the file, downloading it to a specific folder, and then loading it into the software.

TIA.

[jonaswinkler]

Hello, welcome to GitHub!

I want to use paperless to scan in statements. I was hoping it would be able to recognize the statement date and then be able to use that date when renaming the file. for example, if I have a US Bank Statement from 2/1/2021, I want the name of the file to read "USB Statement 2-1-21.pdf". I don't want to use the "added" or "created" date because those dates don't match the statement date. Is this possible?

• Why exactly can't you use the created field? This field is intented to store the date at which a document was issued (invoice date, or issue date of the statement), which is not necessarily the date it was created in paperless.

• You can tell paperless to use the created field for naming files (see documentation, advanced topics)

Also, I want to rename the files because I'll be pulling files into my bookkeeping software, and so it'll be much easier to find files by filename than going into paperless, finding the file, downloading it to a specific folder, and then loading it into the software.

See above. Manual renaming is not supported. However, you can tell paperless to derive filenames from correspondents / dates / types / etc.

[BlankSlate56]

I'll see if I can work out what's going here, this looks like it doesn't do the expected thing.

@jonaswinkler if you are still looking into this, I've loaded about 160 files in the last couple of days. These files cover 19 different correspondents and 13 different document types, so it isn't like I'm force-feeding it nearly identical copies of the same invoices. "Created" hasn't pulled the correct date for a single one of them, despite the dates being near the top of the documents probably 90% of the time. I don't mean this as a complaint so much as a report.

I have noticed that maybe 25% of the time, the date that is pulled into "created" is exactly 1 day ahead of the statement/letter/invoice date that I want. Meaning, if the invoice date is 2021/01/01, paperless is pulling the date in as 2020/12/31. This happened frequently enough that it isn't just a coincidence. Perhaps there is something broken in the Create date algorithm? I'm running 1.1.0.

[BlankSlate56]

For what its worth, below is an example of one of the datasets where the created date was pulled in 1 day early. paperless read this as June 8 2020 instead of June 9. Had to make quite a few changes to parse out identifiable data, don't know how much that'll affect your testing.

North Zeeland Family Dentistry
1400 notaroad Drive
South Nantucketato, NM 50630
(057)265-2237 (570)652-2723
June 09, 2020
Account Aging
Current: $0.00
Fred Flintstone 30 Days: $0.00
19604 570th Ave 60 Days: $0.00
Bells, NM 51607 90 Days: $0.00

[shamoon]

I think the 1 day thing is actually a time zone issue because it’s stored in UTC, I’ve seen it before. We still have some work to do on displaying / entering dates in the UI which I think will help.

And agreed the date matching can be inaccurate often, not sure if thats changed from original paperless or not

[jonaswinkler]

Meaning, if the invoice date is 2021/01/01, paperless is pulling the date in as 2020/12/31. This happened frequently enough that it isn't just a coincidence. Perhaps there is something broken in the Create date algorithm? I'm running 1.1.0.

Set your time zone in the configuration files correctly. The default is UTC. Therefore, any dates assigned to new documents on a certain date at midnight (that's what paperless uses) will display one day early when you're in a time zone that's behind UTC (US, for example).

This change is not retroactive. However, configuring your time zone correctly will cause any new documents to receive the correct date.

I don't exactly know why they added time zone support in OG paperless, but it's there, and I haven't changed it for compatibility reasons.

[BlankSlate56]

ok, I'll check out the time zone thing. I know I haven't tweaked that so I probably missed it in initial setup. That'd make sense for why some of the pulls were one day off. Thanks again.

I actually thought manually entering the dates was easier than I was expecting, I was just hoping that I wouldn't have to do so much of it. Its still fun watching the AI learn the auto-tags and apply them correctly over a growing range of docs.

[balajeek]

Hi, thanks for the awesome project. I am in the process of setting it up, did some few tests to get a feel of it, then i had the same question as this thread subject says:

I have been doing my document management (somewhat manually) and i use Fujitsu scanner which scans and names the files based on what i highlight on the statement i am scanning. There is also a tool hazel on mac were we can right rules so it will find the dates on a scanned document and name them and move them to the approprite folder. I can continue to do that.

But is there a way to tell paperless to take {created_year} {created_month} from the file name that it consumes rather than form the ocr content, because seeing from this long thread mentions paperless most often does not pick the right date from the document, and i tested a dozen document from archive and all are many years off, a 2012 bank of america statement goes into 2017 folder.

possible?