How does data get into the jail?
Once a Data Loader has completed, there may be data in the jail! From Data Loader Central, drill into the Jailer and check the Jailed Rows count against each table. These are data items which have failed our validation tests - either they are structurally invalid (e.g. dates prior to 1753, which SQL Server is not able to handle) or have failed business rules checks which your system administrators have set up. This table tells us how many items are in the jail for each table.
To examine the contents of the Jail, click on the "View" (glasses) icon. You will be able to see which rows failed the Jailer checks, and the rule which failed.
Drilling in to the Jail Details, we have the opportunity to correct the data. So in this case, we set the Customer Name to "Fred"! But there is a problem - what happens NEXT time the data is loaded? The field will be blank once more, and we'll have to correct it all over again. This problem is solved by the check-box at the bottom - Auto-Correct in Future. Providing your table has a Primary Key, Data Academy can be told to always make this correction.
The Difference Between "Save" and "Parole"
Once we've corrected the data, we have two options. "Save" will save our changes but keep the data row in the jail; whereas "Parole" will release the data back into the staging area, where it will be re-validated before being imported into the warehouse. (The re-validation is very important, because there may be more than one rule applied to the import.) Note that once you have Paroled the data, it will disappear from the Jail.
Once you have completed the "Parole" process, return to the Jailer screen, and click the "Update W/H" (Update Warehouse) button to re-run the import into the warehouse. This will import only paroled data into the warehouse - it will not go back to your source system! It will also re-run all of the Jailer rules to ensure that the data now complies in all respects.