Andrew Finkenstadt wrote:
It would appear that SQLite supports indexing every data type. What
would
the efficienty of the index by if I indexed a binary data column?
Similarly, if I wanted to have a datatype which was the equivalent of
"RAW"
(non-interpreted text), it would appear I want to create a database using
UTF-8 encoding and then just be blissfully innocent about whether my
data is
a valid UTF-8 encoding, or would sqlite try to interpret the data and
complain if the encoding resulted in invalid characters?
I'm looking at needing fast access based on hash values, which I
could, of
course, merely store as hexadecimal character text values, but I'd
prefer to
save both disk and memory by using the raw bytes version of the hash
value.
Andy,
SQLite compares blobs using memcmp() so you should have no problem using
it to store your hash values as blobs and indexing on them. I wouldn't
mess with trying to store invalid utf-8 bytes in a text field.
HTH
Dennis Cote
-----------------------------------------------------------------------------
To unsubscribe, send email to [EMAIL PROTECTED]
-----------------------------------------------------------------------------
