Direct From The Source

Windows Search Service Files and Subfolders Structure

Denny Gursky, Software Development Engineer

Windows Experience Find & Organize Team

The configuration and data files of the Windows Search service are stored by default under %ProgramData%\Microsoft\Search. There are two folders under Search: Config and Data.

%ProgramData%\Microsoft\Search\Config

The only file kept in the Config folder is Msscolmn.txt. This is a configuration file describing human-readable names for the properties associated with documents and corresponding full property specifications and property types.

%ProgramData%\\Microsoft\Search\Data

There are two subfolders under the Data folder: Temp and Applications.

%ProgramData%\\Microsoft\Search\Data\Temp

The Temp subfolder is used by Windows Search for creating temporary files.

%ProgramData%\\Microsoft\Search\Data\Applications

The Applications subfolder contains more subfolders corresponding to the applications in the Windows Search service. In this context, "Applications" is not equivalent to "program"; rather, it is a logical entity bound to a specific property store. The only application supported in Windows 7 and Windows Vista is "Windows," so the only subfolder under Applications is Windows.

%ProgramData%\\Microsoft\Search\Data\Applications\Windows

The Windows subfolder contains the subfolders Config, GatherLogs, and Projects, along with a number of *.edb and MSS*.* files. These are Jet database data files and logs that contain the property store.

%ProgramData%\\Microsoft\Search\Data\Applications\Windows\Config

The Config subfolder is always empty in Windows Search 4.0 and later.

%ProgramData%\\Microsoft\Search\Data\Applications\Windows\GatherLog

The GatherLog subfolder contains the single subfolder Systemlndex, which corresponds to the only catalog supported by the Windows Search service. (See the following Projects folder description for details.)

%ProgramData%\\Microsoft\Search\Data\Applications\Windows\GatherLog \SystemIndex

The Systemlndex subfolder contains a number of Systemlndex.*.Crwl and Systemlndex.*.gthr files. The .Crwl files are log files tracking crawl transaction processing results. The .gthr files contain processing results for the notification transactions.

%ProgramData%\\Microsoft\Search\Data\Applications\Windows\Projects

The Projects subfolder contains subfolders corresponding to the different catalogs (projects). Catalogs are the way to partition the index. The only catalog supported by the Windows Search service is Systemlndex, so this Systemlndex folder is the only subfolder under Projects.

%ProgramData%\\Microsoft\Search\Data\Applications\Windows\Projects \Systemlndex

The Systemlndex subfolder contains the subfolders lndexer, PropMap, and SecStore.

%ProgramData%\\Microsoft\Search\Data\Applications\Windows\Projects \Systemlndex\PropMap

The PropMap subfolder contains data files of the proprietary database used for mapping full property specifications to internal property identificators.

%ProgramData%\\Microsoft\Search\Data\Applications\Windows\Projects \Systemlndex\SecStore

The SecStore subfolder contains data files of the proprietary database used for keeping access permissions in the form of security IDs (SlDs) for all indexed documents.

%ProgramData%\\Microsoft\Search\Data\Applications\Windows\Projects \Syste mlndex\lndexer

The Indexer subfolder contains only one subfolder: CiFiles.

%ProgramData%\\Microsoft\Search\Data\Applications\Windows\Projects

\SystemIndex\Indexer\CiFiles

The CiFiles subfolder contains the full text index files themselves, including shadow indexes and the master index. These index files include:

■ The SETTINGS.DIA file, which contains diacritic settings.

■ The *.ci files, which are index files containing indexed words, occurrence information, and references to the documents containing these words.

■ The *.dir files, which are index directory files containing lookup tables for the *.ci files' content to enable fast positioning inside the index without scanning the index file from the very beginning.

■ The *.wid and *.wsb files, which are the fresh test, meaning a table of the documents specifying which information is up to date in the corresponding *.ci file. The trick is to write every *.ci file only once when it is created and never modify it later. If the information in the *.ci file concerning some document is no longer valid, the indexer just marks the document as invalid for this particular *.ci file.

■ The files INDEX.000, INDEX.001, INDEX.002, which implement transactional persistent storage for the index table, which is keeping records concerning all index files in use.

■ The CiMG*.* files, which are merge progress logs that enable the index merge process to continue when interrupted by service shutdown (or even a crash) without having to restart indexing from the very beginning.

■ The CiAD*.* and CiAB*.* files, which are average document length logs that are used for relevance metric calculations.

Was this article helpful?

+1 0

Post a comment