If the images you're storing are larger than the defined page size for the database (which is most likely the case) then you can get better performance and reduced memory consumption by storing the images in the file system and store only paths to the files in the database. This means reading the large amount of data directly from the file system instead of from sqlite's linked-list of pages and bypassing the page caching layer (which you probably don't want for images anyways) and freeing up more of the page cache for real database data.
HTH, Sam On Mon, Feb 11, 2008 at 10:29 PM, Norman Young <[EMAIL PROTECTED]> wrote: > The documentation outlines the C interface for incremental blob > input/output, and mentions the C typedef for a blob handle. > > http://www.sqlite.org/c3ref/blob_open.html > http://www.sqlite.org/c3ref/blob.html > > typedef struct sqlite3_blob sqlite3_blob; > > > Can this same interface be accessed in Python? > > My application manipulates image and audio content. However, on my > constrained platform (Python 2.5, sqlite3, Maemo, Nokia Internet Tablet), > reading and writing the media data via the conventional blob interface (as > illustrated in teh following www.initd.org SnippetsBlobs.py example) could > consume excessive memory. Specifically, all of the binary data are read > into > blobdata object at once, via the read() call. Since the media files can be > arbitrarily large, this call could easily exceed available memory. > > *blob*data = *open*('c:\\*sqlite*3\\img.jpg','rb').read() > > con = *sqlite*.connect(':memory:') > cur = con.cursor() > > cur.execute("Create table picture_table(images)") > cur.execute("Insert into picture_table(images) values > (?)",(*sqlite*.Binary(*blob*data),)) > con.commit() > > > Instead, I need to incrementally read and write media data to a blob, to > avoid consuming arbitrary memory. > > Can this be done from Python via the sqlite3 module? Can you point me to > examples? > > Thanks. > > Norm. > > www.nbyoung.com > www.personalsyndication.com > _______________________________________________ > sqlite-users mailing list > sqlite-users@sqlite.org > http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users > _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
