mikehawkes
Hi again

I'm still having to manually rebuild the database and then re-index it every few days. I'll give an example of what goes wrong - I have a set of board reports that I update every time I get an updated copy. I delete the attachments (there's usually a few of them - usually XLS and DOC files) from the containing thought; then drag the new files from the attachment bar on outlook; the files get added to the thought but fail to index. BrokenFiles.dat contains the names of the updated files and, sometimes, others too.

Doing an index rebuild can take 20 minutes or so ... and I originally reported this problem back in December. I'm wasting a lot of time nursing PB through index rebuilds every time I update a handful of documents.

If there's no fix for this, is there any way that I can schedule a database and index rebuild rather than having to look to see if files exist in brokenfiles.dat and then manually kick off these processes?

Thanks
Mike

PB5026
Vista/XP Pro
Quote
Harlan
Hi Mike,

If the files are getting added to BrokenFiles.dat, that means they cannot be indexed. You should not attempt to rebuild the index as doing so will not accomplish anything - they will not be indexed the second time through either... When items are added to this list, it mean that the process of converting them to text failed and it will fail again if attempted in the future - that is why we store the list, so we don't keep trying to do something we know will not work.

Unfortunately not all files can be successfully converted to text. We are investigating other means of doing the conversions for these cases as part of a future version.
Regards,
-Harlan
Quote
Harlan
Also, note that the search index is completely separate from the database - you should not be rebuilding the database unless you are getting database errors and PB suggests that you rebuild the database.
Regards,
-Harlan
Quote
mikehawkes
Hmm - Ok, so why does it work after I exit PB, delete the brokenfiles.dat file, restart, rebuild? After that, PB finds the files and reindexes successfully (with nothing left in the brokenfiles.dat file afterwards). These are Excel spreadsheets, so nothing particularly weird about the file format.
Quote
mikehawkes
We must be talking at cross-purposes here as I have a test case to demonstrate exactly what happens. Please find a zip file attached.

I created a new brain called test and closed it. I then deleted output and brokenfiles data files in the PersonalBrain directory (to make sure I had a completely clean starting point).

Started PB.

Sent myself an email containing a PDF document containing the word 'conflict'.

Received the email in Outlook.

Opened the email.

Dragged the PDF from the attachment bar in outlook to a free area under the only thought in the 'test' brain.

Attempted a search on the word 'conflict'. Gave a few minutes and tried again. Nothing found in the search.

Took a copy of brokenfiles and output log at this point and put them in the 'test - info' folder. Closed PB. Copied the brain and folder into 'test - info' so you have a copy of that too.

Opened PB, rebuilt the index.

Searched on the word 'conflict'. Found it first time. Brokenfiles.dat empty!

Created a brainzip in case you want to compare with the copy taken above.

Conclusion: On my system, PB does not index files dragged from outlook.

Please help, it's driving me nuts and costs me around 30-45 mins per working day as I have to rebuild indexes every time I add new documents.
Quote
Moe
Hi Mike,  Thank you for taking the time to describe your situation and send us a zip containing further information on your issue.  I conducted a test using the same file "20090131 NLP Practitioner V2.0-U.pdf" which contained the word "conflict" in it and PB was able to find it without me having to re-index.  I noticed in my situation I had the option set in Preferences > Search > Index File Attachments (Checked).  Can you make sure that this setting is turned on.  Also, can you make sure you are running PersonalBrain with Indexing libraries installed.  You can check this by going to Help > About PersonalBrain.

Best regards,
Moe


Quote
mikehawkes
Hi Moe

Thanks for the reply. Index file attachments is checked; About states: "Version 5.0.2.6 with Indexing Libraries. Feb 24, 2009, J-1.6.0_11.". In January I was sent a link to re-install the indexing support but it didn't change the situation.

Thanks again
Mike
Quote
mikehawkes
Hi again

Well, the indexing broke again - this time, I noticed I had a collection of error logs too - I attach a couple for your information.
Quote
Moe
Hi Mike, Thank you for the post. I will have our engineers take a closer look at your error logs to see what can be causing this.

Best regards,
Moe
TheBrain Technologies

Quote
mikehawkes
Great, thanks.
Quote
Moe
Mike, The indexing code has been updated for the next release of PersonalBrain which should fix this.

Best regards,
Moe

Quote
mikehawkes
Excellent stuff. Please pass my thanks on to the team.
Quote

Newsletter Signup  Newsletter        Visit TheBrain Blog   Blog       Follow us on Twitter   Twitter       Like Us on Facebook   Facebook         Watch Us on Youtube  YouTube       

TheBrain Mind Map & Mindmapping Software     Download TheBrain Mind Mapping Software