Bob Please don’t feel like your question about Hive was ignored. It is a very valid question that took some time for us to come to terms with ;) Yes indeed we are thinking about it as we have similar problem with JMS providers. The core issue here is the class path of all that is required by the driver and indeed Hive requires a lot. There is an issue https://issues.apache.org/jira/browse/NIFI-981 where this is already been discussed. We are also considering a raw ORC sink instead of interacting with Hive via JDBC etc. The bad news; As of right now we don’t have much to offer The good news; We certainly plan to have that in 0.5.0 and I am hoping that even before 0.5 will have something in the snapshot that you can try.
So, you won’t need to wait for 1.0, but you may need to wait a little bit ;) Cheers Oleg On Dec 1, 2015, at 6:30 PM, Bob ZHao <bob.zhao.w...@live.com<mailto:bob.zhao.w...@live.com>> wrote: Hi Mark, Thanks for the workaround. This ETL trick always works! Only not sure for SQL injection. Do you have any idea about "configure the DBCP connection pool for hive" ? It's painful and we can't wait until 1.0 release. Thanks, Bob Mark Payne wrote You can also potentially use the ReplaceText processor to build a SQL statement using the Expression Language, something like: INSERT INTO MyTable (col1, col2) VALUES ('${attr1}', '${attr2}') -Mark On Dec 1, 2015, at 5:10 PM, Bryan Bende < bbende@ > wrote: Bob, I think this is slightly different than what you are suggesting, but in the next release there is a new AttributesToJson processor makes it very easy to construct a new json document from values that were extracted using EvaluateJsonPath. Would this makes things a bit easier for sending json to ConvertJsonToSql? -Bryan On Tue, Dec 1, 2015 at 4:14 PM, Bob ZHao < bob.zhao.work@ > wrote: Hello, I like the idea to support more on tabular data target. But it might take a long way to go. I have a quick enhancement suggestion. *COnvertJSONToSQL* only support flat JSON structure instead of the complicated format -- this make sense. But since we can use *EvaluateJsonPath* to filter our the target fields, could we combine them together for a new one *PutPropertyToSQL*? This could be very quick implementation. NIFI needn't to worry about the complex of Json, user can use JsonPath to chop the fields and fragments for customization. Then put the result to target table. BTW, please use a trick to save the raw Json - by pointing to a column. Thanks, Bob -- View this message in context: http://apache-nifi-developer-list.39713.n7.nabble.com/PutSQL-feature-enhancement-PutPropertyToSQL-tp5536.html Sent from the Apache NiFi Developer List mailing list archive at Nabble.com. -- View this message in context: http://apache-nifi-developer-list.39713.n7.nabble.com/PutSQL-feature-enhancement-PutPropertyToSQL-tp5536p5541.html Sent from the Apache NiFi Developer List mailing list archive at Nabble.com<http://Nabble.com>.
