Define data checking to specify how PowerExchange handles record mapping errors and field-level data errors during an extraction.
The data checking options are:
Record mapping errors. You can specify that PowerExchange stops the extraction or that it continues after skipping the row and writing a message to the log file. Default is to stop the extraction.
If you do not want to take the default action of stopping the extraction when a mapping error occurs, you can select
skip row and write log file msg
. Informatica recommends that you put any necessary effort into mapping the data by using tools such as record IDs, so that PowerExchange skips rows only as the last resort.
Data errors in fields. You can specify that PowerExchange stops the extraction or that it continues after performing one of the following actions:
Skipping the row and writing a message to the log file.
Setting the field to null.
Replacing the bad data with values according to the
Data Defaults
tab of the
Data Map Properties
datalog box. This option is available for DATE, TIME, CHAR, VARCHAR, and CLOB field types.
Default is to stop the extraction.
Selecting field types on the
Data Defaults
tab in the
Data Map Properties
dialog box provides a default setting if you select the
Use Data Map rules
option on the
Field Checking
tab of the
Field Properties
dialog box. Selecting field types also defines whether the fields that result from calls to user-defined functions are checked. Because it is seldom a good idea to perform data checking on user-defined fields, Informatica makes the following recommendations:
On the
Data Defaults
tab of the
Data Map Properties
dialog box, do not select field types.
On the
Field Checking
tab of the
Field Properties
dialog box, avoid selecting
Use data map rules
and define rules on a field-by-field basis.
Because performing data checking increases the CPU cost on the machine where the mapping runs, Informatica recommends that you perform data checking only when absolutely necessary.
Data checking is required only when fields contain data that cannot be loaded onto the target.
Invalid data in DATE or TIME columns often causes problems for target systems and might need to be checked.
Most modern relational databases are tolerant of CHAR, VARCHAR and CLOB columns that contain control characters with hexadecimal values less than the value of the space character. However, some legacy databases which cannot tolerate these control characters.