22021 Character Not In Repertoire

When a client sends or receives text data to the server, the following error may occur:

ERROR: invalid byte sequence for encoding "UTF8": 0xb2

Possible causes:

• The encoding of the client system environment does not match the encoding parameter of the database client (e.g. cp936 for the client system environment and UTF8 for the database client).
• The encoding of the executed SQL script file does not match the encoding parameter of the database client.
• the text stream contains binary data like NUL bytes that are not allowed within a string.

Solution:

On Linux, you can use locale to view the encoding of the client system environment:

$ locale -k charmap
charmap="UTF-8"

On Linux, you can use file to view the encoding of the SQL script file:

$ file --mime test.sql
test.sql: text/plain; charset=utf-8

On Windows, you can use chcp to view the encoding of the client system environment:

> chcp
Active code page: 936

The code page 936 shown above is an alias for the GBK encoding.

You can view the database client encoding as follows:

postgres=# SHOW client_encoding;
 client_encoding
-----------------
 GBK
(1 row)

This converts the text data from client-side encoding to database server-side encoding.

You can view the current server encoding like this:

postgres=# SHOW server_encoding;
 server_encoding
-----------------
 UTF8
(1 row)

If the encoding parameter of the database client does not match the encoding of the client system environment, you can modify the database client encoding as follows:

SET client_encoding = 'UTF8'

If the encoding of the SQL script file does not match the encoding parameter of the database client, you can modify the encoding of the SQL script file as follows:

$ iconv --from-code=GBK --to-code=UTF-8 test-gbk.sql > test-utf8.sql