Few answers: > MSSQL three part naming: Right now in Drill a particular connection will need to be focused on a particular database in MSSQL. If you need to connect to multiple databases, you'll need to configure multiple instance of the JDBC storage plugin.
1. Is there anything I can do to up the logging and so be able to track > these kinds of issues myself and maybe contribute fixes back? > Best two things: 1) Make sure to enable verbose errors via: ALTER SYSTEM SET `exec.errors.verbose` = true; 2) Install Logback Lilith (https://github.com/huxi/lilith) and enable the socket appender by uncommenting these blocks in your logback file: https://github.com/apache/drill/blob/master/distribution/src/resources/logback.xml#L19 https://github.com/apache/drill/blob/master/distribution/src/resources/logback.xml#L80 2. Im trialling using some fairly exotic types in postgres, mainly arrays > and hstore. I'm assuming drill won't support those out of the box and I'll > need to use views to flatten them? If I myself wanted to contribute support > back where would I start? > Two main pieces of code. The first is in the Calcite Project which Drill uses. It is here: https://github.com/apache/calcite/blob/master/core/src/main/java/org/apache/calcite/adapter/jdbc/JdbcSchema.java We would need to make sure that this correctly presents the data types desired as part of the table schema. The second thing would be to address the reading behavior in the Drill RecordReader in this code: https://github.com/apache/drill/blob/master/contrib/storage-jdbc/src/main/java/org/apache/drill/exec/store/jdbc/JdbcRecordReader.java Would be happy to help in any way. Thanks, Jacques > Cheers, > Kieran > On 6 Nov 2015 7:49 pm, "Jacques Nadeau" <[email protected]> wrote: > > > He Kieran, > > > > There were some issues with the 1.2 JDBC storage plugin. It had issues > with > > certain data types including TEXT/BLOB. We're in the process of releasing > > version 1.3 of Drill which should solve many of these issues. Can you try > > the current release candidate of Drill 1.3 to see if that solves the > issues > > you are seeing? > > > > You can download the release candidate here: > > > > http://people.apache.org/~jacques/apache-drill-1.3.0.rc0/ > > > > thanks, > > Jacques > > > > -- > > Jacques Nadeau > > CTO and Co-Founder, Dremio > > > > On Fri, Nov 6, 2015 at 10:09 AM, Kieran Benton <[email protected]> > > wrote: > > > > > Hi all, > > > I'm having a whale of a time getting the JDBC storage support working > > with > > > Drill 1.2 :) Could someone please give me a pointer at what I'm doing > > > wrong, or even just confirm to me that the feature isnt baked yet and > has > > > issues? > > > > > > I've done the following (I'm on Windows for my sins): > > > > > > 1. Installed Drill in embedded mode > > > > > > 2. Added the postgres jdbc driver to the jars / thirdparty folder as > > > referenced at > > > > > > > > > http://stackoverflow.com/questions/33311547/unable-to-configure-postgresql-jdbc-driver-using-apache-drill-1-2 > > > > > > 3. ./sqlline.bat > > > > > > 4. !connect jdbc:drill:zk=local (admin/admin) > > > > > > 5. Accessed the web console and added a new storage 'customers' with > the > > > definition: > > > { > > > "type": "jdbc", > > > "driver": "org.postgresql.Driver", > > > "url": "jdbc:postgresql://localhost:5432/customers", > > > "username": "customers", > > > "password": "customers", > > > "enabled": true > > > } > > > > > > 6. Success! (took a while to get this far, finding the 'right' driver > > wasnt > > > easy) > > > > > > 7. Try 'SELECT * FROM customers.profile;' in the console -> > > > > > > Error: VALIDATION ERROR: java.lang.NullPointerException > > > [Error Id: 62cb35b1-db49-4740-ac38-4d748a080f3a on 172.31.9.168:31010] > > > (state=,code=0) > > > > > > > > > > > > And then I'm completely stuck. Where do I go from here? > > > > > > Cheers, > > > Kieran > > > > > >
