As of NiFi 1.2.0, after GetFile you can use the PutDatabaseRecord processor with a CSVReader, giving it a schema if you know the layout, or if the CSV file has a header row you can get the column names from that using "Use String Fields From Header" for the Schema Access Strategy property. Prior to NiFi 1.2.0, after GetFile you probably want a SplitText to get each row in its own flow file. Do you know the number of columns for the CSV file? If so then you can use a regex in ExtractText; assuming there were four columns you might have a dynamic property called "column" set to something like: ([^,]+),([^,]+),([^,]+),([^,]+) That should give you attributes "column.1", "column.2", "column.3", and "column.4". Then you can use ReplaceText to generate a SQL statement, perhaps something like: INSERT INTO myTable VALUES(${column.1},${column.2},${column.3},${column.4}) Then send that to PutSQL. Another option (prior to NiFi 1.2.0) is to convert the CSV to Avro (see this HCC article), then ConvertAvroToJSON, then ConvertJSONToSQL, then PutSQL.
... View more
