No announcement yet.

Beyond Compare 4 comparing 2 pdf files yielded incorrect mismatch

  • Filter
  • Time
  • Show
Clear All
new posts

  • Beyond Compare 4 comparing 2 pdf files yielded incorrect mismatch

    I compared 2 pdf files, each 643 pages long, appx 3MB in original file size.

    Ran Beyond Compare 4 on my Windows 10 PC and it yielded an incorrect mismatch. That is, it reported the 1st pdf file had a value 2.5 in 1 column and value 0.1 in that same column for the 2nd pdf file.

    However, reviewing the original pdf files confirmed no difference and further, the actual value in both pdf files was 0.3

    Comparing pdf files for similar output, (i.e., same report using different parameters,) yielded the expected 100% successful compares.

    So far, only thing to note is that the 1st pdf file was reported by Beyond Compare 4 to have 3,230,969 bytes whereas the 2nd pdf file was 3,201,219 bytes. There were differences in file size for the similar output pdf files, too, so that may be a red herring.

    And so, at this point, we're going with this is a bug, unless someone has any ideas as to what may be the cause.

    Thanks in advance for any suggestions.

  • #2

    BC4's Text Compare performs a conversion to plain text in the background. This shouldn't alter any values, but the text can be aligned unexpectedly depending on inserted line breaks, which can present as differences. Was the value changed, or alignment altered?

    If you use Adobe itself to perform a File menu -> Save As .Txt for each file, and compare Adobe's output in Beyond Compare, how does it appear here?

    One alternate conversion parameter which can help as a different alignment, is to clone the existing Adobe File Format, and replace the External Conversion -> Command Line with:
    pdftotext.exe -enc UTF-8 -table -nopgbrk %s %t
    Aaron P Scooter Software


    • #3
      Thanks, Aaron, for your fast response. Your point "text can be aligned unexpectedly" looks relevant as while the row flagged as mismatching was actually matching (in the original pdf files), it turns out the mismatch was the data two rows above the one flagged that was actually mismatching (again, looking in the original pdf files directly / just visually, i.e., outside of Beyond Compare).

      Also appreciate your citing an alternate conversion parameter and we may pursue that, too. For now, it's simply good that the actual mismatch was identified.