We are using Informatica as ETL and it is installed on AIX 7.2 server. We have installed 16.10 TD ODBC driver on the same. We are using this driver to connect to Teradata and load the data. We have chosen UTF8 as the character set in our odbc.ini file. Also the connections in Informatica are also created using UTF8 as the Code Page. In our Informatica code, we are able to read Chinese characters from a source file and load them successfully into a target file too. But while loading the data from the same source file to Teradata, we see an issue. We are just wondering if there is an issue with the TD ODBC driver on AIX.
Has anyone faced any similar issues?
Thanks in advance
What issue do you see?
The target column should be defined with CHARACTER SET UNICODE.
And you will need to use a compatible client character set (e.g. UTF8) when connecting to query the database also.
We have defined the target columns as Unicode. We are able to insert chinese characters thru direct Insert sqls from SQL Assistant. But only while loading thru ETL tools using AIX ODBC driver, we see this issue. Also now we are encountering the issue while reading the data from TD using the same driver. The relavent contents in odbc.ini file are as below:
We have followed the same steps as what we normally do while connecting from ETL tools installed on Linux.. But on AIX, we are seeing this behaviour.
Let us know if any additional settings need to be made.
To add more details, we have created an ODBC connection rather than Relational connection for connecting to Teradata.
Any more settings to be done in odbc.ini file?
Those ODBC settings seem OK. When you insert / query via SQL Assistant, is that connection using UTF8 session character set also?
Sometimes comparing in hexadecimal form a small amount of input data against the resulting database values will help understand what is happening.
e.g. use od -x inputfile in Unix / Linux / AIX and CHAR2HEXINT(targetcolumn) function in Teradata.