Operations that targets the items in a data-log
The job is designed so that you should find the logs you're looking for, or a wider search.
Followed by a filter-stage where you may filter out the more precise items you need to operate on.
The operations that can be performed on data log items are performed on the results after find- and filter- stages.
In the store step you can control for example if the changes should be persisted or just saved as resource in the current workflow context.
Examples for the filters:
* LeftSide: item.metadata.val1
* Operator: ==
* RightSide: hejsan
* LeftSide: [rndnum(1,10)]
* Operator: >
* RightSide: 7
* LeftSide: item.phone
* Operator: ==
* RightSide: incomingmessage.sender
* LeftSide: metadata.incomingmessage-keyword
* Operator: contains
* RightSide: item.metadata.key
This job can be connected to the following workflow elements.
- Jobs
- Fields marked with
*require input. - You need to ad the prefix
metadata.to access data in your WFC
When working with datalog items, you may store searchable metadata for each individual log item by having the key prefixed with _search_example (underscore search underscore data-name). So the example "_search_example" set to "test123" will store a searchable value "example=test123". Finding data log items by metadata can utilize this, compose a search phrase with the key (excluding the underscore-search-underscore part and separate key and value with equal sign = )