This post is a session which i did yesterday at philly force admin group. My session covered topics on how to use jitterbit data loader, a live demo on jitterbit and also best practices to avoid major problems on data load.
Here is synopsis.
To prevent problems , before you do a data load, you need to do the following
- Check date fields and their formats to match salesforce date format
- If you use excel, watch out for truncation of fields like zip code for leading zero’s.
- Plan to verify data loads using reports and views in salesforce to show the count of records created or modified.
- If your input file is more than 20,000 records, split large input files to multiple small files to avoid data locks
- Check email fields for valid formats
- Use 18 digit Salesforce Ids for updates/upserts
- If there are picklist fields, those fields should match salesforce data picklists.
- Ensure that there is a data backup prior to load.
- Make sure that you turn off validation rules, workflow rules for bigger data load
- Create an id field or a field to identify source of the file import.
After Data Load, you need to do the following
- Test with a salesforce report on count of records inserted or updated.
- On bad data loads, have a roll back plan to wipe out bad records.
- Check for random samples on data load.
- Run demand tools or data quality tools to check for duplicates.
- Run an impact analysis before deleting bad data .
- Verify rollup fields on parent object which are dependant on child objects on most recent child record created or modified.
Using the above tips, you can prevent major data load issues. Please feel free to download my presentation below . I would appreciate if you can post your comments on your horror stories on data load failures and feel free to post your comments on any questions you have. Please click like if you like my post or feel free to share it with your friends.
Click below to download presentation.