Faced the following problem: inserting some binary data from the App to "bytea" field in our PostgreSQL 9.2 DB, we found out, that its length is doubled. The App is written in C++ and uses SQL API library to access Postgres. Reference to official doc: http://www.sqlapi.com/HowTo/blobs.html
Saying that data length is doubled means, that before insert command int the app we have: sContent.length() = 19 meanwhile in postgres we get: select length(bindata) = 38, bit_length( bindata ) = 304 from binpacket
In C++ App we do:
SAString sContent = SomeFunctionConvertsByteArrayToSAString(bindata);
cmd.Param("bindata").setAsLargeBinary() = sContent; //SA_dtLongBinary <=> BYTEA
//watch: sContent.length() = 19
cmd.Execute();
and have field length doubled in postgres (select length(bindata) = 38 ).
I wrote some simple Python3 script, that inserts binary data from file to the database:
bindataStream = open('C:\\Temp\\bin.dat', 'rb').read()
cursor.execute("INSERT INTO binpacket( bindata ) VALUES (%s)",
(psycopg2.Binary(bindataStream ),))
and the length of the data inserted is equal to the source binary data file.
That leads me to the conclusion, that it is possible to insert binary data to Postgres 9 without having its length(size) doubled. But is it possible when the data source is a bytearray?
Please, help to find an correct solution how to insert bytea binary data to Postgres. The explanations of the situation described would be appreciated too.
Best regards, Anthony
To fix the described Issue, update to SQLAPI++ 4.1.2 (it is free for legal customers, like we are). The developers fixed the porblem with binary data for Postgres 9.x. in SQLAPI++ 4.1.2 Release.