Best Of

Have you looked at the Split Column tile in Magic ETL?

You used to be able to do this with the CLI tool and specifically the backup command; however, it's since been deprecated per the documentation but you may be able to still utilize this

> help backup

backup: Performs a backup operation, to AWS S3, for all Domo entities of type <t>.

  backup -k <S3 Key> -s <S3 Secret> -b <S3 Bucket> -p <S3 Path> -r <S3 Region> -t <Backup Type>

  usage:
    -a,--role-arn <role-arn>                            Role ARN
    -b,--bucket <bucket>                                AWS S3 bucket
    -cd,--card-dataset <card-dataset>                   Card Dataset Id
    -cda,--card-dataset-auth <card-dataset-auth>        Card Dataset Auth
    -cdd,--card-dataset-domain <card-dataset-domain>    Card Dataset Domain
    -d,--detail-dataset <detail-dataset>                Report Dataset
    -e,--external-id <external-id>                      Role External Id
    -k,--key <key>                                      AWS S3 key
    -m,--kms <kms>                                      KMS Key
    -n,--notifications <notifications>                  Notification List
    -p,--path <path>                                    AWS S3 path
    -r,--region <region>                                AWS S3 region
    -s,--secret <secret>                                AWS S3 secret
    -t,--type <type>                                    Backup Type
    -x,--summary-dataset <summary-dataset>              Report Dataset

Alternatively, you can use the undocumented APIs to replicate the dataflow or utilize a third-party python package like https://pypi.org/project/domolibrary/

In addition, and perhaps this is what you meant by manual work, you can create copies of a dataflow with one click. I often do this if I can going to rip apart a dataflow as a backup.

@TC1199A and I met and determined that the data was likely being removed from a filter tile in the ETL. He will rebuild and if that wasn't the case report back here. While walking through the ETL I shared some things that Taylor and Davis found helpful and thought I would share them here for those that don't know:

1. When using snapshots/recursive dataflows there is a chance for data loss. It is always a good idea to export data before making changes that could potentially be destructive so that you have a backup.
2. You can revert to previous versions of a dataflow in the history tab as shown below:

You could use the start date of the week as your date range so that you have a single week value like your week number but is an actual date. You can use a formula tile and a formula like:

`date` - INTERVAL (DAYOFWEEK(`date`)+1) DAY

Then you can group your data on this date to either take the min, max, sum or whatever aggreagation you need for the week.

You can also just click on the cog next to the cast data type in the Alter Columns tile and under Bad Values, choose Null. The tile will then convert all the values to integer that are actually integers and then null out all the values that don't convert.