+1 ExecuteScript does allow incoming flow files... also if your scripting language is Groovy you can use Sshoogr to execute remote commands, see my example here. ... View more

I confirmed this to be a bug in ConvertJSONToSQL, I have written up NIFI-4071, please see the Jira for details. ... View more

Or try adding an extra slash before your path (see here for explanation): file:///post/postgresql-42.1.1.jre7.jar ... View more

That error is a bit different, can you share the associated stack trace from the log(s) (logs/nifi-app.log, e.g.)? ... View more

Do you mind putting this into a new HCC question? Although the problems may be somewhat similar, you will likely get answers that don't apply to the original question here, so to avoid confusion, I think a new question is the right way to go 🙂 Thanks in advance! ... View more

If you want to use a PreparedStatement (which can improve performance if PutSQL operates on a number of records up to the specified Batch Size), then your SQL could be: insert into Table(X, Y, Z) values(?,?,?) To use this you'd have to set up your attributes as follows: sql.args.1.value=<value of X> sql.args.1.type=<type of X's value, see here for numerical values of datatypes>sql.args.2.value=<value of Y>sql.args.2.type=<type of Y's value>sql.args.3.value=<value of Z>sql.args.3.type=<type of Z's value> If you know the types of the columns, you can do all this work with a single UpdateAttribute processor before the PutSQL processor. For example, if X and Y were integers and Z was a string/varchar: sql.args.1.value=${X}sql.args.1.type=4sql.args.2.value=${Y}sql.args.2.type=4sql.args.3.value=${Z}sql.args.3.type=12 See the PutSQL documentation (especially the "Reads Attributes" section) for more information. If you don't know the types or order of your columns, then I recommend @Eyad Garelnabi's solution instead. ... View more

It looks like you are trying to connect via the Zookeeper port, but there are some issues with this: You'll likely have trouble because of the issue described in NIFI-2575. Even if that were not the issue, the following two things would have to be done. I believe you'd have to set some variables in the URL (such as serviceDiscoveryMode=zooKeeper;zooKeeperNamespace=hiveserver2) The Zookeeper port is not exposed via the sandbox by default, you would have to forward that port if using NAT. For these reasons, I recommend you connect via the standard HiveServer2 port (10000), which is exposed by default for the sandbox's NAT configuration. ... View more

Are you using the JDBC Driver for HAWQ? If so, you should be able to use PutSQL to execute your INSERT...SELECT statement as-is. If you have tried this, what error(s) are you getting or did you have any trouble configuring things? ... View more

True, if you are just inserting a single new field at the top-level of a flat JSON doc, this should work fine. However you may want an optional newline between the array and the first object if the JSON is pretty printed. Also, would this work if you had a nested array of objects further down in the file? I guess you could put in a start-of-line into the regex, but if the JSON has weird whitespacing you may run into the same problem. I recommend JOLT in general because it handles that kind of thing, but if you know what your input JSON looks like, this solution is simple and works well. ... View more