Hello, I'm cooking up some blob reading code based on the following example ->
http://www.sqlite.org/cvstrac/wiki?p=BlobExample The author prepares a statement and reads the blob value in the following manner -> *if*( rc==SQLITE_ROW ){** *pnBlob = sqlite3_column_bytes<http://www.sqlite.org/capi3ref.html#sqlite3_column_bytes> (pStmt, 0); *pzBlob = (unsigned char *)malloc(*pnBlob); memcpy(*pzBlob, sqlite3_column_blob<http://www.sqlite.org/capi3ref.html#sqlite3_column_blob> (pStmt, 0), *pnBlob); } ... rc = sqlite3_finalize<http://www.sqlite.org/capi3ref.html#sqlite3_finalize> (pStmt); At the point prior to the call to sqlite3_finalize, the blob is going to more stored twice in memory. In my application, however, memory is very tight. Is there an easy way I can get around sqlite3_finalize freeing the blob data? Deferring the finalize until I am done processing the blob really isn't a palatable option.... On a similar note, are there any hacks available to write a blob to a file using a limited memory buffer? As far as I can tell, the sqlite API requires that a blob be read in its entirety, which is rather unfortunate when the blob is bigger than memory... Thanks, -Stan
