Detects duplicate files regardless of name, size in Salesforce. This is not a package - please customize as you see fit. This is also not for small issues of duplicates; this is to handle a case of runaway triggers, invalid process builder, or bad workflow rules that caused a swelling of attachments.
You can create the following objects & upload the code to your own Salesforce instance. After uploading, simply queue the process by calling the following from either a controller or anonymous execution window. Make sure to set your own key and secret first at the top of the FileScanner class
// src/classes/FileScanner.cls
String key = 'PUT_YOUR_OWN_KEY_HERE';
String secret = 'PUT_YOUR_OWN_SECRET_HERE';Then queue the FileScanner class. After that is done, queue up the FileScannerCleanup class.
// Do initial scan
Id batchJobId = Database.executeBatch(new FileScanner(2019, 'myemail@mydomain.com'), 500);
// After this is done, run the cleanup operation
Id batchJobId = Database.executeBatch(new FileScannerCleanup(10, 'myemail@mydomain.com'), 500);
// After cleanup results, we can run the marker to locate the first instance of a given file that was discovered
Id batchJobId = Database.executeBatch(new FileScannerMarker('myemail@mydomain.com'), 10);
// If this fails or is interrupted by governor limits being hit, queue a rollback operation here to mark them again for processing
Id batchJobId = Database.executeBatch(new FileScannerMarkerRollback('myemail@mydomain.com'), 10);
// After marking, we can process the files marked for deletion DO NOT DO THIS WITHOUT BACKING UP FILES
Id batchJobId = Database.executeBatch(new FileScannerDestroyer('myemail@mydomain.com'), 500);The code works by reading the blob of the attachment body and using an MD5 hash to generate a unique signature for the file. If you'd like a stronger algorithm, you can consult the Documentation Here
// src/classes/FileScanner.cls
Blob hashBlob = Crypto.generateDigest('MD5', attachment.Body);
String hash = EncodingUtil.convertToHex(hashBlob);After hashing the file it will create a new object (as we cannot make fields on Notes, Attachments, etc) that contains information on the file. After computing the hash of all files, you will want to queue a cleanup operation to remove files under a given threshold.
We recommend using a report on the included object and it will automatically have "open file" so you can preview results.
Included in this is a batch operation specifically aimed to mark files for deletion. The code will look for the first known instance of a hash and mark it for saving while marking all other files with the same hash for deletion.
Due to various conditions such as file sizes, amount of records, etc we created a Rollback class to roll back the marking process if it is interrupted. We cannot automatically queue this, so please reference the processing apex jobs to know if it encountered a governor limit.
This final component will delete files marked by the file marker and set the Deleted__c checkbox to true. Do not run this until you have made a backup.
We found removing any copy of the hash after the first known version (file created date) worked best. Keep one, remove the others.
Rememeber to take backups!
MIT