More matching fields for Rules

ts57_expensifyts57_expensify Expensify Customer Posts: 2

Expensify team, you software is already doing OCR or PDF to text already to pull the merchants name from the receipt and all the other data. Why did you stop at just allowing rules to match the merchants name? The rule engine is very limited, please consider making it more powerful

I would like to have a cascade of match fields so I can move receipts into reports based on other values in the receipts. Consider adding simple logic and matches (case insensitive)

Example below:

Match Merchant "Name" (this is done already)

====Add more below=======

Choose logic (and, or, not)

Second Term "text data to match on"

Choose logic (and, or, not)

Third Term "text data to match on"

   Choose logic (and, or, not)

Forth Term "text data to match on"

**** Would really be cool if you could allow regEX and just have user select between simple or full regEX. You are using a regEX under the covers anyway, just allow the user to specify the full expression!

This would be simple to add I believe based on my background as a software creator for many years.

Please consider this feature request.

The use case I have:

I get 100's of receipts per week from homeDepot, on 34+ projects. Each receipt has a bit of text that says the project name (other places do the same, or the buyer adds the name to the receipt before submit; example "Palm", "21st", "Bishop". I want to sort each receipt into a report unique to the job. I do this manually, but it is painful and this is something a computer is uniquely suited for -- repetitive task based on known input and outcomes!



1 votes

Understanding the Issue · Last Updated


  • Rachael HopkinsRachael Hopkins Expensify Success Coach - Admin Posts: 780 Expensify Team

    Thanks for your Idea @ts57_expensify - don't forget to vote!

    Right now, SmartScan only 'reads' the Date, Merchant and Amount. Are you envisaging with this Idea that there would be additional data points picked up? Can you elaborate on these data points?

  • ts57_expensifyts57_expensify Expensify Customer Posts: 2

    I am asking that you scan the document fully into a one text blob. Then execute the match rules supplied by the user. There would be no additional datapoint picked out, just the ability to move (process by other things?) the receipt to a report based on the match rules. I am asking that you allow the user to use REGEX rule set to get fine grain matches done.

    But since you bring it up, REGEX does have the ability to capture the matched text and put it in variables that can be accessed programmatically. If you can do that and have a general purpose capture and placement in report details that would be a nice add. Use case - on receipt it says in hand written form "Client:Bob" and a match/capture sequence found that. It could then "Bob" could be placed into the comment of the receipt "Dinner with <match1> client". This would go a long way to helping automate the receipt processing. You could also use this to set currencies and other data in the report and receipt.

    To me it just seems the processing engine is already there for most of this, just the UI is missing. Or, that at least that how I would have built the system, get big text blob then start applying RegEX to find data. So if that is the way the system is built, it is just a matter of allowing RegEX added by client to make the system more powerful.


  • Rachael HopkinsRachael Hopkins Expensify Success Coach - Admin Posts: 780 Expensify Team

    That's for the detail @ts57_expensify I will leave this open for voting! Personally, I'm not completely au fait with how our data parsers work, but I do know that we create them ourselves. I don't believe that they work how you assume, so I think this might be a major piece of development.

Sign In or Register to comment.