@smartraman  This can also be accomplished through a different and more complex configuration of the ReplaceText processor: Using below input content example: { "TOT_NET_AMT" : "[\"55.00\"]", "H_OBJECT" : "File", "H_GROSS_AMNT" : "[\"55.00,58.00\"]", "TOT_TAX_AMT" : "[9.55]" }   I would set up the replaceText processor as follows: Instead of just searching for those character patterns and replacing them with nothing, I break entire input line-by-line in to a series of capture groups.  That way I can omit the capture groups matching the patterns you want removed ([ or [\" or \"] or ]) and then manipulate the capture group containing a possible comma separated list, so that only the last value in that list is returned. I used below java regular expression which results in 5 capture groups: (.*?)([\Q[\E]\\\"|[\Q[\E])(.*?)(\\\"[\Q]\E]|[\Q]\E])(.*?)$ I then used the following Replacement Value in which I used NiFi expression language against the 3rd capture group.  If that capture group does not contain any commas, the entire string is returned.   With example above and this configuration, you end up with the following new content: { "TOT_NET_AMT" : "55.00", "H_OBJECT" : "File", "H_GROSS_AMNT" : "58.00", "TOT_TAX_AMT" : "9.55" }   If you found this helped with your latest query, please take a moment to login and click on "Accept as Solution" below this response. Thank you, Matt       ... View more

@mickt  I looks as though you may be hitting: https://issues.apache.org/jira/browse/NIFI-8938 This issue was introduced via a change that was introduced as part of NiFi 1.14: https://issues.apache.org/jira/browse/NIFI-8723 If you found this response addressed your query, please take a moment to login and click "Accept as Solution". Thank you, Matt ... View more

@midee    Not a lot of details in your question here, but the exception: SunCertPathBuilderException: unable to find valid certification path to requested target is telling you that the trust could not be determined in the TLS handshake that occurred between client and server. Essentially this means that yoru truststore did not contain all the necessary TrustedCertEntries. The complete trust chain must exist in the truststore. Lets say you have Certificate "cert-X" signed by Certificate Authority (CA) "CA-A". Owner: CN=cert-X, OU=test Issuer: CN=CA-A, OU=CA-intermediate Then that CA "CA-A" was signed by another CA "CA-root" Owner: CN=CA-A, OU=CA-intermediate Issuer: CN=CA-root, OU=CA-root Then the CA "CA-root" is signed by itself: Owner: CN=CA-root, OU=CA-root Issuer: CN=CA-root, OU=CA-root So in order to trust the certificate CN=cert-X, OU=test, the truststore would need to contain the complete trust chain meaning it would need to have a TrustedCertEntry for both "CA-A" and "CA-root" There may even be more CAs in that trust chain.  You need every public cert for each ca all the way to the root CA (owner and issuer the same) to have complete trust chain. A mutual TLS handshake would require trust in both directions. Clients certificate must be trusted by server and server's certificate must be trusted by client. In a 1-way TLS handshake you only need trust in one direction. Client must be able to trust server's certificate only since client would not be sending a certificate to the server. If you found this response addressed yoru query, please take a moment to login and click "Accept as Solution". Thank you, Matt ... View more

@smartraman  You can use a ReplaceText processor to remove these special characters from your json. Using your example, I could produce yoru desired output using the following java regular expression:   (\\")|[\Q[\E]|[\Q]\E]   Source: My ReplaceText processor was configured as follows:   result: If you found this response addressed your query, please take a moment to login and click on "Accept as Solution". Thank you, Matt ... View more

@midee    The condition NEL statement and resulting actions NEL statements will be unique for each rule you create. For example: Rule 1: Conditions:    ${attr1:trim():length():minus(${attr1:trim():replaceAll(' ',''):length()}):plus(1):equals(1)}   Actions:   ${attr1:substring(0,1)}   Rule2: Conditions:   ${attr1:trim():length():minus(${attr1:trim():replaceAll(' ',''):length()}):plus(1):equals(2)}   Actions:   ${attr1:substring(0,1)}${attr1:getDelimitedField(2,' '):substring(0,1)}   Rule3: Conditions:   ${attr1:trim():length():minus(${attr1:trim():replaceAll(' ',''):length()}):plus(1):equals(3)}   Actions:   ${attr1:substring(0,1)}${attr1:getDelimitedField(2,' '):substring(0,1)}${attr1:getDelimitedField(3,' '):substring(0,1)}     Rule4: Conditions:   ${attr1:trim():length():minus(${attr1:trim():replaceAll(' ',''):length()}):plus(1):equals(4)}   Actions:   ${attr1:substring(0,1)}${attr1:getDelimitedField(2,' '):substring(0,1)}${attr1:getDelimitedField(3,' '):substring(0,1)}${attr1:getDelimitedField(4,' '):substring(0,1)}   You will notice in the Conditions NEL statement, the last number is incrementing for each rule and in the Actions NEL statement the statement gets a bit longer with each additional string  If you found these answers to your query were useful, please take a moment to login and click "Accept as Solution" on all answers that helped. Hope this helps, Matt ... View more