This question has been asked before (here) but never answered so I'm hopeful that today is my lucky day and someone knows the answer. Because unfortunately, the documentation does not cover the topic.
In the UpdateAttribute there is an option to Store State of an attribute by setting the value to "Store state locally". But to do that I first need to set the "Stateful Variables Initial Value", but the documentation doesn't say how to do that, and everything I've tried fails. The documentation provides an example for referencing an existing attribute variable, theCount, but there is no mention of how to initialize this attribute variable.
I've attempted setting the initial value using:
- theCount=1
- {"theCount":1}
- ${theCount=1}
- 1
- ${setStateValue('theCount')} <- grasping at straws with this one
None of these work. When I look at the value of the attribute theCount it always just shows: "Empty string set".
I'm including a screen shot of my processor so you can see the configuration.
Also, including a screenshot of the Attributes tab showing the theCount attribute.
This worked for me :
UpdateAttribute Configuration:
Notice the "Stateful Variables Initial Value" is set to 1
The UpdateAttribute success relationship has the result flowfile with the following attribute:
If I run the flow again , the new flowfile will have the following attribute:
Its strange that It did not work for you when set the initial value to 1 ? I dont see anything else wrong. If that still did not work for you, can you try to upgrade Nifi just to see if its not a bug with 1.19.1. I use 1.20 for testing but you can upgrade to the latest.
If you find this helpful please accept solution.
Thanks
Regarding the state being persistent, actually you can clear the state by right click on the processor then select "View state" and then click on the "Clear state" link:
This should reset the state to the initial value.
Regarding your second question of being able to initialize more than one variable, you can define as many stateful variables as you need in one processor, however they all can have one initial value. If you are looking to define different initial values for different stateful variables then you have to create a different UpdateAttribute processors for the variables that have common initial value. Another option - I never tried - if you want to utilize one processor then you can use Advanced option to define rules that will set different initial value based on a common conidition, however you have to be careful how you set the first value on the first flowfile. For example, if you have two stateful variables Attr1 & Attr2 where the first flowfile has Attr1 = 0 & Attr2 = 1 then increment both afterward, you can define the UpdateAttribute as follows :
Notice how I set the Initial Value to Empty String since its required to set some value when using stateful.
Under Advanced I defined two rules: one to initialize Attr1 to 0 and another to initialize Attr2 to 1 when each is set to empty string :
Rule for Attr1:
Rule for Attr2:
Make sure to set the FlowFile Policy top right to "use original" , otherwise it will duplicate flowfiles for each matched rules if "clone" is used.
When setting the same attribute in Advanced mode and in basic mode, the Advanced will take precedence if rules are met, so the first time the increment wont run.
The first flowfile will have the initial values since the rules are satisfied.
The second flowfile the rules are not satisfied therefore the increment will happen:
Depending on what you want to see on the first flowfile then you can adjust the initial values accordingly. Not sure if this will work for all scenarios but you can try, otherwise use different processor as I stated above. Also if anyone thinks that goes against best practices or might cause problem please advise.
If you find this helpful please accept solution.
Thanks
This worked for me :
UpdateAttribute Configuration:
Notice the "Stateful Variables Initial Value" is set to 1
The UpdateAttribute success relationship has the result flowfile with the following attribute:
If I run the flow again , the new flowfile will have the following attribute:
Its strange that It did not work for you when set the initial value to 1 ? I dont see anything else wrong. If that still did not work for you, can you try to upgrade Nifi just to see if its not a bug with 1.19.1. I use 1.20 for testing but you can upgrade to the latest.
If you find this helpful please accept solution.
Thanks
Regarding the state being persistent, actually you can clear the state by right click on the processor then select "View state" and then click on the "Clear state" link:
This should reset the state to the initial value.
Regarding your second question of being able to initialize more than one variable, you can define as many stateful variables as you need in one processor, however they all can have one initial value. If you are looking to define different initial values for different stateful variables then you have to create a different UpdateAttribute processors for the variables that have common initial value. Another option - I never tried - if you want to utilize one processor then you can use Advanced option to define rules that will set different initial value based on a common conidition, however you have to be careful how you set the first value on the first flowfile. For example, if you have two stateful variables Attr1 & Attr2 where the first flowfile has Attr1 = 0 & Attr2 = 1 then increment both afterward, you can define the UpdateAttribute as follows :
Notice how I set the Initial Value to Empty String since its required to set some value when using stateful.
Under Advanced I defined two rules: one to initialize Attr1 to 0 and another to initialize Attr2 to 1 when each is set to empty string :
Rule for Attr1:
Rule for Attr2:
Make sure to set the FlowFile Policy top right to "use original" , otherwise it will duplicate flowfiles for each matched rules if "clone" is used.
When setting the same attribute in Advanced mode and in basic mode, the Advanced will take precedence if rules are met, so the first time the increment wont run.
The first flowfile will have the initial values since the rules are satisfied.
The second flowfile the rules are not satisfied therefore the increment will happen:
Depending on what you want to see on the first flowfile then you can adjust the initial values accordingly. Not sure if this will work for all scenarios but you can try, otherwise use different processor as I stated above. Also if anyone thinks that goes against best practices or might cause problem please advise.
If you find this helpful please accept solution.
Thanks
