Content matching; Date not matched

I'm using the content matching on a PDF that has been OCRed.

So I'm trying to Date match on "Date: " followed by a date.

If I just do a custom date match on the date I'm OK, it all works, but that's too general- I don't want it triggering on dates elsewhere.

If I put "Date" followed by the 'Anything' element followed by the same custom date match then it doesn't work.

I selected the text and pasted it into Text Edit and it shows as "Date:21 November 2012"

I ran "mdimport -d2" on the file and it shows the text as "Date: 21 November 2012"

Either of those should work OK with my scheme. Does Hazel not get the same text as mdimport? Is there a way to discern what the text is that Hazel is seeing?

Thanks in advance.

The text Hazel gets is slightly different and may be in a different order. If you want to dig into it, you can look in Hazel's bundle for a hazel importer program. Run that on the file (redirect the output to a file as it may not display properly in Terminal) and you can see the text as Hazel would. Or, you can email support and send in the file and I can take a look at it for you.

Thanks! That has really helped. I've got my matches working a little better now.

One match that I can't seem to make work is on the text (extracted using Hazel);

Code: Select all: 439194 settlements@share.co.uk 15 November 2012 SM

In order to try and match this I use;

Code: Select all: share.co.uk(anything)(Date Match)

And for (Date Match) I use (31)(space)(December)(space)(1999)

Does the (anything) token not match carriage returns?

Thanks in advance. Oh, and Merry Christmas!

(anything) does not match line separators as that could result in very bad things (like your whole document sucked into that one variable). A single space will though.

Mr_Noodle wrote:If you want to dig into it, you can look in Hazel's bundle for a hazel importer program. Run that on the file (redirect the output to a file as it may not display properly in Terminal) and you can see the text as Hazel would.

I am having a lot of trouble matching dates and other text that should be very simple.

I would like to try looking at Hazel Importer. I found the app in the Hazel package contents.

Can you please explain in more detail how to use it? Specifically, how do I "redirect the output to a file"?

Thank you.

bd_e wrote:
Mr_Noodle wrote:If you want to dig into it, you can look in Hazel's bundle for a hazel importer program. Run that on the file (redirect the output to a file as it may not display properly in Terminal) and you can see the text as Hazel would.

I am having a lot of trouble matching dates and other text that should be very simple.

I would like to try looking at Hazel Importer. I found the app in the Hazel package contents.

Can you please explain in more detail how to use it? Specifically, how do I "redirect the output to a file"?

Thank you.

I'd also a like more detailed explanation on how to use the hazel importer program. This sounds like it will be very helpful in creating and debugging content matching rules.

Thanks

It requires some knowledge of the command line but in short, it should be run as follows:

(path to hazelimporter) (path to the file you want to process) > (path to the file you want to write the output to)

Where you supply the appropriate paths as indicated above. Do not include any of the parentheses.

Thank you -- this was helpful. I was able to generate the output file, but I am still having trouble matching.

In this case, the output file includes this section of text:

Page1 of 3
0080 32
$187.95 $190.77

I want to capture the first dollar amount.

I created a custom token with the pattern "(123).(123)". (I'm creating a custom token because I want to use the dollar amount later in the actions that change the file name.)

I created a match for "Contents // contain match // $[custom token]".

When I click the preview eye and then the attributes window, nothing has matched for my custom token.

How can it be that there is no match?

I suggest emailing support. Include your rule as well as a sample file indicating the problem.