[ https://issues.apache.org/jira/browse/NIFI-4940?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16388109#comment-16388109 ]
ASF GitHub Bot commented on NIFI-4940: -------------------------------------- Github user zenfenan commented on a diff in the pull request: https://github.com/apache/nifi/pull/2513#discussion_r172590173 --- Diff: nifi-nar-bundles/nifi-standard-services/nifi-record-serialization-services-bundle/nifi-record-serialization-services/src/main/java/org/apache/nifi/csv/CSVHeaderSchemaStrategy.java --- @@ -79,4 +81,12 @@ public RecordSchema getSchema(Map<String, String> variables, final InputStream c public Set<SchemaField> getSuppliedSchemaFields() { return schemaFields; } + + private String normalizeNameForAvro(String inputName) { + String normalizedName = inputName.replaceAll("[^A-Za-z0-9_]", "_"); --- End diff -- How about this `[^A-Za-z0-9_]+`? So rather than adding N underscores for continuous non-compatible characters, it adds a single `_`. > CSVReader - Header schema strategy - normalize names > ---------------------------------------------------- > > Key: NIFI-4940 > URL: https://issues.apache.org/jira/browse/NIFI-4940 > Project: Apache NiFi > Issue Type: Improvement > Components: Extensions > Reporter: Pierre Villard > Assignee: Pierre Villard > Priority: Major > > When using the CSV Reader with the Header Schema Strategy access, we should > add a property (boolean) in case field names should be normalized to be avro > compatible. Otherwise it won't be possible, for instance, to use the > UpdateRecord processor and access a field with a "weird" name. -- This message was sent by Atlassian JIRA (v7.6.3#76005)