I'm using WOA pdf-Excel to extract values from a table. An example of a line in the table would be, say:

itemname  12,123  23,456  34,789

using lookbehind, I can easily extract the numbers using the regex (?<=itemname\s+(\d+,\d+\s+){n})\d+,\d+    where n=0,1,2 to extract the 3 numbers.

But, there's no lookbehind support in WOA pdf-Excel, so (as a very inexperienced newbie) I need some expert advice as to how to do this - I've been hugely unsuccessful to date!

Thanks to anyone who can assist.
 

what do u need the lookbehind for? the regex

\d+,\d+

will extract numbers like

12,123

from input 

12,123  23,456  34,789

just fine. Please explain again what the problem is.

Sorry, obviously didn't explain well enough.

The pdf table described consists of many lines of data such as the one described. The first column describes what the data is ( "itemname" in my example) and each row of data has a different itemname. I am only interested in extracting the data from a single row with the itemname I need. Also, I need to extract that same row from the same table in about 100 pdf's in a batch, so extracting everything and sorting it out using excel would be messy. I need to do this every month. It's very important that the process is automated.

 In short, I want a regex to extract just the row of data named "itemname" and I want each number in the row in a separate excel column.

 Hope this makes it clearer- and that someone can help!
 

i' m not familiar with the regex engine u r using., so i'll show you generic approach to the problem as I see it. To iextract N numbers after the field name *itemname* u can use

(?<=itemname)(\s+\d+,\d+)+

against your sample input:

itemname  12,123  23,456  34,789

the numbrs will be captured as N capturing groups of your match:

  12,123  23,456  34,789

u'll have to programmatically access all the groups in your Match object, and then Trim() leading white spaces from the strings and them Cast() then to numeric datatype

12,123 

23,456 

34,789

If you dont want to Trim(), u can use this regex that directly will capture the numerics

(?<=itemname)(?:\s+(\d+,\d+))+

but for that to be done your regex engine must support non-capturing groups (?:  ....)

You can try something like this:

 itemname\s+(?:(\d+,\d+)\s*)*

 I don't know what degree of control you have on your groups or matches afterwards.

That add-in does not support ?<= but seems to support other pattern constructs, however you might best ask the author how to use it: dwilkie@gmail.com

---

This is exactly my problem: the regex engine does not support the (?<=....) construct. I need a way to simulate it.

I would then recommend that you just use the app to get the lines of data and use regex witihin Excel to split the line data (in a single column) into multiple columns.

If in your app if you use a pattern such as:

itemname.*?(?=\r|$)

you should end use with a column A with cells that look like:

itemname    10,000  12,222  8,999

Then you could construct a macro to select column A that would read the cells and run this regex routine against them to split them into B,C,D, etc.:

Sub RegExp_Late()
'Late binding
'Dimension the RegExp objects
    Dim RegEx As Object, RegMatchCollection As Object, RegMatch As Object
    Dim C As Range, i As Long

    ' create the RegExp Object with late binding
    Set RegEx = CreateObject("vbscript.regexp")

    ' set the RegExp parameters
    With RegEx
        'look for global matches
        .Global = True
        .Pattern = "[^ ]+"
    End With

    ' set the Excel range to parse the seelection
    For Each C In Selection
         i=0
        Set RegMatchCollection = RegEx.Execute(C.Value)
        For Each RegMatch In RegMatchCollection
            i = i + 1
            C.Offset(0, i) = RegMatch
        Next
    Next

    Set RegMatchCollection = Nothing
    Set RegEx = Nothing
End Sub

That should result in:

A1:itemname    10,000  12,222  8,999

B1:itemname

C1:10,000

D1:12,222

E1:8,999

If you need additional assistance with Excel macros that would be out of the scope of this forum, I provided the macro code as an example of what I would do.

---

many thanks for this - looks like I'll need to do it in 2 stages, as suggested.