SNAPSHOT ETL DATAFLOW

WorldWarHulk

Hi Beatiful People - 

 

I was able to learn how to create a snapshot etl dataflow,

 

just a quick question,

where in the dataflow can I add a column that shows the date for the historical data and the date column for the appended data?

 

 

WorldWarHulk

yes. or can I use the batch date for that?

jaeW_at_Onyx

Ask your CSM if it's possible to pair Salesforce with UPSERT or APPEND.

https://www.domo.com/appstore/connector/salesforce-with-upsert-connector/overview

 

While you can use ETL to build out this pipeline, Salesforce + UPSERT would be the Domo-preferred method (as opposed to building a recursive dataflow) for building history assuming the dataset is small enough (low millions of rows) and with some updates to MagicETL, this will scale pretty nicely.

GrantSmith

Hi @WorldWarHulk 

 

It appears that the screenshot you posted is from the Knowledge base article (https://knowledge.domo.com/Prepare/DataFlow_Tips_and_Tricks/Creating_a_Recursive%2F%2FSnapshot_ETL_DataFlow) which is what I'm going off of. The way that's configured is it joins the historical with the live data and ignores / filters any historical data which exists in the live file thus you lose and information from the historical dataset.

 

Are you wanting to keep a date of when it was first added to the dataset and another when it was last updated on the dataset?

 

WorldWarHulk

yes. or can I use the batch date for that?

WorldWarHulk

you mean to say, it will lose any data when the new data is appended? 

 

I need a data to be appended and create a snapshot of historical data...

 

my data comes from salesforce which doesnt store historical data... 

jaeW_at_Onyx

Ask your CSM if it's possible to pair Salesforce with UPSERT or APPEND.

https://www.domo.com/appstore/connector/salesforce-with-upsert-connector/overview

 

While you can use ETL to build out this pipeline, Salesforce + UPSERT would be the Domo-preferred method (as opposed to building a recursive dataflow) for building history assuming the dataset is small enough (low millions of rows) and with some updates to MagicETL, this will scale pretty nicely.

WorldWarHulk

subscribed to your youtube channel Onyx Reporting

user09742

I'm appending daily data for historical use. Because of the size I don't want store all of the dates. I just want the last 8 days only and remove the other dates. How can I do this? I can filter for the last 8 days, but it still stores other dates data.