Windows 7: Search does not find files that it should, directory is indexed, and *.php included on the extension list

When I tried detailed searching on my 32-bit PC, it didn't work and rebuilding the index was extremely slow.  So I reinstalled Windows 7, and among the first things I did then was to install Adobe Reader (for the PDF IFilters) and set up search indexing, relocating the index to C:\WindowsSearchIndex. I  specified indexing of D:\Martin, below which is all my data (documents, music, pictures, videos).  (Because of this separation of OS and data, reinstallation is not such a big deal; I use the redirection property of My Documents, etc.)

I also went into the Local Group Policy Editor (gpedit.msc, Windows Pro and Ult only I think, although you can also do this by editing the registry):  Local Computer Policy/Computer Configuration/Administrative Templates/Windows Components/Search, and enabled "Disable indexer backoff" to speed up the indexing.

I let it index overnight, and today Search finds PDFs by name and content, .rtf files likewise, just as I detailed above.  As soon as I finish putting the basic system in place, with Microsoft Office etc., I'm going to back up the system partition with EaseUS ToDo Backup, then I can revert to it if things start to deteriorate again.

So Windows Search seems to work fine on a fresh installation, but somehow goes bad subsequently.  Problem (sort of) solved.

P.S.  Installed MS Office, now Search finds e-mails by content.   Also searching by content of .php files works.    Q.E.D.

Hi,

I have just conducted the experiment, similar to what I did above for PDFs, of constructing several .rtf files (each containing a line or two of Shakespeare) and moving them to different folders under "My Documents\Office".  I then immediately searched for them from Explorer at the "My Documents" level, not really expecting to find them straight away.  However, Windows appears to have indexed them almost instantly.  Not only could I find them both by filename and by content, but when I corrected one (from "Unleash..." to "Let slip the dogs of war"), that was instantly locatable by content too.

 I searched using terms like:

 breach  [which was in a filename]

 slip type:rtf   [in the contents]

 heaven ext:.rtf  [in the contents]

 or just:

 ext:.rtf

 which latter found them all.

 So there is something odd going on here, perhaps to do with some corruption that occurs the longer Windows has been installed.  Rebuilding the index might help, although I know some people have tried it without success.

 I agree with you about Microsoft; they are no help, although the amount they charge for the OS, they certainly should be.  I suspect from the various fora I have looked at that MS actually don't have a clue either.  All I can recommend is reinstalling the operating system (ouch) and setting up and trying the indexing straight away.  If it works, back up the system as a system image immediately.  Reloading it is not such a terrible task if you keep your data on a separate partition.

Regards,

J M Ward

Just occurred to me: have you gone into Services.msc and checked that the Windows Search service has actually been started?  It should be Automatic, but may have got disabled somehow.

As with many software/hardware problems, Microsoft is ignoring this issue, and assuming that the problems are mostly human error.  Many of the suggested corrections don't work or miss the point altogether.

Here is a very simple manifestation of the problem.  One of my folders has a slew of "rtf" files.  Yes, the folder is in my indexed locations, and yes, Windows search is set to index and search the contents of "rtf" files.  No, I was not using McAfee when I conducted this experiment.  I am using the latest version of Windows 7 64-bit, all up-to-date.

Nothing I did could coax or cajole Windows into finding a simple text string in the contents of an rtf file.

I have read through this entire thread, and nothing I see hear begins to fixes the problem, except possibly the suggestions of 3rd party searches.

It would be a refreshing change if Microsoft actually took note of this topic, acknowledged that there is, in fact, a problem, and addressed it.  I would be shocked if that happened; it would be uncharacteristic for them.

-ETC

Dear Paul,

I thought your problem interesting, so I have done the following experiment:

Running Windows 7 Ultimate 64-bit, I made sure Adobe’s 64-bit IFilter plugin was installed, and that the PDF filters were set for contents indexing (Indexing Options/Advanced Options).  Then, in Advanced Options, I selected Rebuild to rebuild the index from scratch (incidentally, I always create a directory C:\WindowsSearchIndex and relocate the index there, using the same dialog box).

Finally, I prepared three PDF files, “2007 Alpha paper.pdf”, “2007 Beta paper.pdf”, and “2007 Gamma paper.pdf”, which I put two levels deep inside “My Documents”.

Overnight, Windows completed indexing at 496,436 items, which I reckon is about right.  Today I started searching from the “My Documents” level.

1)       When I entered “2007”, Windows took around 35 seconds to find my three files plus all other files with “2007” indexed  in the content.

2)       I specified “2007 type:pdf”, and it took around 15s to find the three PDFs and all other PDFs containing “2007”.

3)       Specifying “2007 ext:.pdf”, it took 10s to find files as in (2).

4)       Then I specified ‘System.FileName:~<”2007” ’, and the three corresponding PDFs (only) were found within three seconds.  (The syntax means “find files with name starting…”).

You can find search syntax such as in (4) here:

http://windows.microsoft.com/en-gb/windows7/advanced-tips-for-searching-in-windows

http://windows.microsoft.com/en-gb/windows7/advanced-tips-for-searching-in-windows

5)       Finally, I used the 64-bit version of Agent Ransack (see my post just above) to look for “2007*.pdf”, from the “My Documents” level as above, and it found all three in 35s.

6)       Finally finally, I used the Advanced Search in Adobe Reader to look for “2007”.  This took a long time, because AR does a full un-indexed content search for the word specified.  The search appears to be of content only, and not of filenames, so it only found those of my prepared files which actually had “2007” in the text.

So all I can suggest to you is to try different search syntax (see reference above and my reference to a “Windows Search” document in my post above yours), and if that doesn’t work, try rebuilding your search index overnight.

Alternatively, use Agent Ransack, or (gulp) reinstall Windows to the stage that you have it how you basically want it, then make a system backup (use EaseUS, not Windows).  I run a relatively fresh installation of W7U, which I refresh from an external system backup if things seem to be going belly-up.

Of course, you may be running 32-bit Windows, which could for some obscure reason behave differently.  I am in the process of doing the same tests on a 32-bit machine, but for some reason it has been taking forever to index.  I have found a way of speeding up the indexing, and I’ll post when I have a result.

Incidentally, if you have the opportunity to put “tags” on your PDF files when creating them, that will also speed up using Windows Search to find them.  See references above.

J M Ward