MS SQL FTS Index Population Error

Hi,

try the following to resolve the issue...

Download and install the Adobe PDF iFilter from Adobe, and test it using the iFiltTst on a directory with PDF files. It does index the pdf files.

I think Adobe has recently released IFilter 6.0

Then try to reindex them in SQL server...

If you still face the issue... run the following...

use master
go
sp_fulltext_service 'load_os_resources',1
go
sp_fulltext_service 'verfiy_signature', 0
go
reconfigure with override

let me know if u face any issues...


all the best...

Hi,
I have the Adobe PDF Filter 6.0 installed and correctly indexing 700+ PDF's.
It is that filter (as i see it) that is raising an error with some of the PDF's.

I am trying to locate which PDF's are having the issue, and then what the issue is with them.

Thanks anyway.

Locating the pdf's causing issues is hard to find...

check the following issues...

·         Verify that the PDF file is not damaged or password protected. The PDF iFilter cannot index PDF files that cannot be opened in Acrobat. Open the PDF file in Adobe Reader 6, Adobe Acrobat 6 Standard or Adobe Acrobat 6 Professional to verify that the file is not damaged or password protected.

·         Verify that the PDF file does not contain Security Permissions prohibiting you from searching the file.  Open the file and choose File > Document Properties > Security.  Verify that the value for "Content Copying or Extraction" is "Allowed". If it is "Not Allowed", the PDF iFilter cannot index the text.

·         Verify that the PDF file contains searchable text. You can verify if there is searchable text in the PDF file by opening it in Reader or Acrobat and selecting the text with the Text Select tool.  Or, you can open the file and using the Edit > Select All command to select text.  If neither method highlights any text, it is likely that the PDF contains a source image of text which the PDF iFilter cannot index.  You can then use a tool like Acrobats Paper Capture tool (Document > Paper Capture > Start Capture&) to convert the source image of text into searchable text which PDF iFilter can index.

·         Verify that the PDF file contains text that is properly encoded.  If the file contains searchable text, yet the Acrobat Search tool (Edit > Search) cannot find the text, then the text may not be properly encoded.  Select a word using the Text Select tool, copy it to your clipboard, and paste it into a text editor (e.g. Notepad).  If the word is not legible in the text editor, the file contains text or fonts that are not properly encoded.

·         Ensure the PDF file has a ".pdf" filename extension. Clients, such as Index Services, find the PDF iFilter by looking up the filename's extension in the Windows Registry.

·         If the original document was created in a Windows 95 application and uses TrueType fonts, either reformat the text with Type 1 fonts or use Adobe PDF Printer to recreate the PDF file. In Windows 95, PDF files generated from PostScript files containing TrueType fonts include text that is not searchable.  Both the Microsoft and Adobe PostScript® printer drivers remap TrueType fonts at PostScript generation time, which disconnects the character you see on screen from that character in the font (for more information, see http://www.adobe.com/support/techdocs/12f6e.htm ).

Excellent, that's what I can do once I locate which of the 700+ Pdf's the Filter is having trouble with...

The PDF's are only in the database,  so to test them I would have to extract them from the database first and some are over 200mb..

I need to know how to locate the specific PDF files causing the issue with the indexing by translating the error message...I am happy to extract and correct/ replace those.

Thanks
TheTimp

Ok now we are getting somewhere..

the 'a' didn't work but I found a common word for most projects , but it gave me some results - brilliant!. (See the attached code)
It exposed some PDFs that had been secured. (I can fix that) :-)

But it also found a PDF that is not secured and I can search with acrobat reader that contains the <common word>, and returns results inside of Acrobat...perhaps i will get the PDF re-created from the original document and see if the problem persists?. (FYI Properties say it was created with Distiller 8.1.0 , PDF version  1.4  for Acrobat 5.x)

But getting back to my question how do the logs point to those files..It seems the ID's in the logs are not too helpful?
The full-text key value in the log is not in my list of PDFFilesID's.... is it not an ID.. how does it relate?

Thanks
TheTimp

Select cast(PDFFilesID as NvarChar(50)),PDFName from
PDFFiles
Where
PDFFilesID not in 
(
SELECT FT_TBL.PDFFilesID   
FROM dbo.PDFFiles AS FT_TBL
INNER JOIN FREETEXTTABLE(PDFFiles, *, '<Common Word>') AS KEY_TBL
       
        ON FT_TBL.PDFFilesID = KEY_TBL.[KEY])
Order By cast(PDFFilesID as NvarChar(50));

Select allOpen in new window

I think its the ID generated by full text while creating the log....

OK,

So there is no relation between the ID's in the Log and the table.
How disappointing.

You provided a Solution to the issue so I will considier this issue closed.

Thanks
TheTimp