In DB Scheme docs, LCR table, priority should be int

But this is ok, as char is an integer data type and it is translated correctly :
for mysql -> TINYINT
for postgres -> SMALLINT
for oracle -> NUMBER

I guess this is ok, right ?

Regards,
Bogdan

I'm afraid that it makes troubles. I found this because it didn't worked for me with MSSQL via ODBC. char(4) was translated to 4 digits number and every lcr_reload priority was changing. I changed it to tinyint and everything became OK and lcr_dump gives me correct rules.

I think it just typo in DOCs. In MySQL installation scripts the data type is int.

But in both docs (http://www.opensips.org/html/docs/db/db-schema-devel.html#GEN-DB-LCR) and db schema, the priority is defined as "unsigned char" (with no size).

So, for MSSQL, this "unsigned char" is translated to a 4 digits number (xxxx) ? if so, what is the problem? do you have priorities larger then 4 digits, or the db_unixodbc is not properly loading the value?

In /scripts/mysql/lcr-create.sql in 1.4.4:
priority TINYINT UNSIGNED NOT NULL
oracle script:
priority NUMBER(5)
postgres script:
priority SMALLINT NOT NULL

Only the doc says that it's char.

Yes it's translated to 4 digits number (xxxx).
The problem with ODBC and MSSQL is that this 4 digits number (xxxx) is every time different for every lcr_reload. This translation is somehow not unique. And I have in DB only number from 1 to 16 in priority column.

Changing it to tinyint fixing the problem.

To avoid any issues with less standard data-types, I already changed th column type to int.

Thanks and regards,
Bogdan