Handling LONG/BLOB data
l What makes LONG / BLOB / MEMO data special?
n Not practical to pre-allocate fixed size buffers for worst case
l Fetching LONGs - treat as normal fields after setting:
n $dbh->{LongReadLen} - buffer size to allocate for expected data
n $dbh->{LongTruncOk} - should truncating-to-fit be allowed
l Inserting LONGs
n The limitations of string literals
n The benefits of placeholders
l Chunking / Piecewise processing not supported
n So you're limited to available memory
n Some drivers support an unofficial blob_read() method