Problems with spaces in extracted texts

[cew]

Hi,

I have a big problem using text extraction.
Problem 1: I get words with spaces between each character
Problem 2: Words are enreadable because the charcters are shuffelt

In the appending file you can see the original pdf and the resulting text file and the result created by Adobes Acrobat.

I already tried out setting the PXP_TETextComposeOptions MergeDistanceX. Not better...

What's wrong?

Best regards
cew3

[Lzcat - Tracker Supp]

Problem is because your file contain fonts with non-standard encodings, so xcpro cannot correctly determine characters widths (especially space width) and therefore cannot compose letters correctly (yes, in this document each letter is displayed separately, so text composing is not so simple).
I'm affraid that I cannot say when this can be fixed, however you have possibility to correct this yourself. Please use PXCp_ET_GetElement function instead of PXCp_ET_GetPageContentAsTextW and then compose text as you wish.
HTH.

[cew]

I'm affraid that I cannot say when this can be fixed, however you have possibility to correct this yourself. Please use PXCp_ET_GetElement function instead of PXCp_ET_GetPageContentAsTextW and then compose text as you wish.

How do I use PXCp_ET_GetElement to get the complete text of my document?
I couldn't find any demo c# source where this function is used.

Best regards
cew

[Corwin - Tracker Sup]

Hi cew,

Please check attached example (frmTextExtraction.cs -> btnSimple_Click).

HTH.

[cew]

Please check attached example (frmTextExtraction.cs -> btnSimple_Click).

Thank you for the code example.

But this way is incredible slow:
Extracting an pdf with about 10 mb takes over 7 minutes.
Using PXCp_ET_GetPageContentAsTextW takes about 15 seconds.

Best
cew

[Lzcat - Tracker Supp]

Then you will need to optimize it. This is just a sample, without any optimizations - as I can see there are a lot of memory allocation/deallocation operations (two per each text element), this may be a bottleneck.
Also, xcpro compose text inside, without making copies as PXCp_ET_GetElement does, so your code will be little slower, sorry.