Hi @SushmaNaya , I believe this was asked before here but I dont think there is  a way to do it in one shot. you probably have to do it in custom script like executeScript processor where you can have part of the data stored as an attributes (depending how big the part is ) and then the other part is your flowfile content or have both stored as attributes (not recommended). Once you get the flowfile the execute script will get the content of the flowfile and the other aprt form the attribute construct your multi part body and send it that way. By the way if you are using Nifi 2.0 you probably can utilize python extensions to create your own custom processor using pure python when you can use different packages to help you with that. You can find example here: https://community.cloudera.com/t5/Support-Questions/Nifi-1-18-multipart-form-data-with-binary-part-and-json-part/m-p/386351#M246014 https://www.w3schools.com/python/ref_requests_post.asp Hope that helps ... View more

Hi, Another option is to use FreeFormTextRecrodSetWriter for that. The documentation is lacking about this unfortunately but you can find some examples like this if you google it. All you need is a ConvertRecord processor to get the desired result. here is an example :     - GenerateFlowFile is to simulate generating the csv input: - ConvertRecord: which takes the CSV input using CSVReader and Record Writer using the FreeFormTextRecordSetWriter:   - CSVReader Service Configuration: You can use default configuration.   -FreeFormTextRecordSetWriter: The Text used in the above Service to provide desired output: username = ${username} first name = ${"first name"} middle name = ${"middle name"} last name = ${"last name"}   Output: username = test_user first name = test_FN middle name = test_MN last name = test_LN username = test_user2 first name = test2_FN middle name = test2_MN last name = test2_LN   Hope that helps.               ... View more

Hi, Actually I went against the main  input since the second input (simplified version) is not valid json. Hopefully the following will work:     [ { "operation": "shift", "spec": { "description": "priceRuleName", "startDatetime": "startDate", "endDatetime": "enddate", "locations": { "0": { "location": "storeLocation" } }, "customAttributes": { "*": { "id": { "PR_PRICE_RULE_PROMO": { "@(2,valueBoolean)": "priceRulePromo" } } } }, "tiers": { "*": { "priceRuleAttributes": { "*": { "id": { "PR_BRANCH_NAME": { "@(2,values)": "paths.[&5].pathName" }, "PR_TARGET_SEGMENT": { "@(2,values)": "paths.[&5].customerCondition[#2].targetSegments" } } } }, "rewards": "paths[&1].calculatePrice", "conditions": { "*": { "conditionsMerch": { "*": { "item": { "": { "@(2,excludeInd)": { //group categories under yes\no include ind "0": { "@(4,categoryId)": "paths[&9].catalogCondition.category.yes.categoryList[]" }, "1": { "@(4,categoryId)": "paths[&9].catalogCondition.category.no.categoryList[]" } } }, //there is a valye in item -> products "*": { "@(2,excludeInd)": { //group products under yes\no include ind "0": { "@(4,item)": "paths[&9].catalogCondition.products.yes.productList[]" }, "1": { "@(4,item)": "paths[&9].catalogCondition.products.no.productList[]" } } } } } } } } } } } } , { "operation": "shift", "spec": { "*": "&", "paths": { "*": { "*": "&2[&1].&", "catalogCondition": { "*": { //products or category "*": { //bucket each yes\no group into its own array element "*": "&5[&4].&3.&2[#2].&", "$": "&5[&4].&3.&2[#2].include" } } } } } } } /**/ ] Also something to consider is using JSLT. JSTL is another json transformation language similar to xquery in syntax ( check reference here) . There is jslt processor in nifi for that. In my openion jslt works better in these situation specially when you trying to query data from very nested structure using some complex logic to match on certain values . You probably have to write less lines in jslt and the logic is easier to follow than jolt (of course if you have get familiar with jslt syntax and how it works first to make sense of it) Here is how this transformation looks using jslt: import "http://jslt.schibsted.com/2018/experimental" as exp let priceRulePromoArray = [for (.customAttributes) .valueBoolean if(.id=="PR_PRICE_RULE_PROMO")] let mainInfo = { "priceRuleName":.description, "startDate":.startDatetime, "enddate":.endDatetime, "storeLocation":.locations[0].location, "priceRulePromo":$priceRulePromoArray[0] } let paths= [for(.tiers) { "pathname":flatten([for(.priceRuleAttributes) .values if(.id=="PR_BRANCH_NAME")]), "customerCondition":[for(.priceRuleAttributes).values if(.id=="PR_TARGET_SEGMENT")], "calculatePrice": .rewards, "catalogCondition":[for(.conditions) let groupProductByInd = exp:group-by([for (.conditionsMerch) . if(.item!="")] ,.excludeInd, .item) let groupCategoryByInd = exp:group-by([for (.conditionsMerch) . if(.categoryId!="")] ,.excludeInd , .categoryId) let products = [for ($groupProductByInd ) { "include": if(.key==0) "yes" else "no", "productList":.values } ] let category = [for ($groupCategoryByInd ) { "include": if(.key==0) "yes" else "no", "categoryList":.values } ] if(any($products)) {"products":$products} else {} + if(any($category )) {"category":$category} else {} ] } ] {"paths":$paths}+$mainInfo   ... View more

Not sure why do you need to use the replace text processor in this case. I provided that in my sample above as an example anda a way to simulate getting new data by replacing the original content with something else. You can think of replaceText as if Im doing InvokeHttp and getting different flowfile content in the response releationship. If you got the data from module C just link directly to the join enrichment. As long as you have the correct writer\reader configured for each fork then you should be good. ... View more

It all depends on the complexity of the data you are working with. If you are taking about data transformation (converting to timestamp, replacing quote, ...etc.) then maybe groovy is the way to go. JSLT has some function that can you help you accomplish this as well like string replace and parse-time functons but Im not sure that is everything. Im not sure where did you get the impression that the nifi community doesnt recommend using groovey and if you find an article about that please share. I think its more of an issue with your Boss not wanting you to do any scripting to avoid not having this supported by others than you. The processor is there for you to use. Actually there is a dedicated processor for groovey called ExecuteGroovyScript. I think the ExecuteScript processor might get deprecated since its redundant. The only issue that  I can find that warns about this processor is the fact the script is getting compiled for every flowfile and that might get expensive and impact the performance if you have a big script and working with large data volume. To avoid running into those scenarios, Nifi provides other alternative like InvokeScriptedProcessor (using groovey as well) or develop your custom processor in java (.nar) where the code is compiled once and done. The jslt processor also re compiles the script but it uses caching to avoid having to do that every time. In terms which performs better: groovey or jslt? Im not sure and I have never tested but you can do some stress testing and let us know :). ... View more