Applicable Ephesoft Transact Versions: 4.x or above

This article provides steps on how to configure the key-value pair for KV extraction in the Chinese language at the database level.

• For MySQL/MariaDB:

1. To insert Chinese characters in the key_pattern and value_pattern columns of the kv_extraction table, change the collation of these columns from latin1_swedish_ci to utf8_bin by executing the following MySQL query:

ALTER TABLE `kv_extraction` CHANGE COLUMN `key_pattern` `key_pattern` VARCHAR(700) NULL DEFAULT NULL COLLATE ‘utf8_bin’ AFTER `key_fuzziness`;

ALTER TABLE `kv_extraction` CHANGE COLUMN `value_pattern` `value_pattern` VARCHAR(700) NULL DEFAULT NULL COLLATE ‘utf8_bin’ AFTER `use_existing_key`;

• For MS SQL server:

1. To insert Chinese characters in the key_pattern and value_pattern column of the kv_extraction table, change the data type of these columns from varchar to nvarchar by executing the following  MSSQL queries:

alter table [ephesoft].[dbo].[kv_extraction]

alter column key_pattern nvarchar(700);

alter table [ephesoft].[dbo].[kv_extraction]

alter column value_pattern nvarchar(700);