Alteryx has the ability to connect to data sources using fat clients and ODBC but not JDBC. If the ability to use JDBC could be added to the product it could remove the need to install fat clients.
Yes yes a thousand times yes! Now working with snowflake and I have to say their ODBC driver just doesnt cut the mustard. Write 1 row at a time and is super slow. I had to create a bulk uploader in the end. So much fun and I now use it all the time. But this instead of a standard driver. Non technical people would struggle with this. A better way forward is this.
At this point in time there are still no development plans for the ability to connect to data sources using JDBC on the road map. Should this change and we are able to include it on the road map, we'll be sure to update this idea accordingly.
All, let me, please, write a longer response that I hope will get us into more detailed discussion.
Let me start with a provoking statement: I am the Product Manager for the connectors, and for me having jdbc connectivity in Alteryx Designer does not make much sense.
Alteryx Engine is written in C/C++, as are the ODBC libraries, for us this is the native environment, we are not Java-based
we do not have JVM part of our install; adding JVM installation would significantly bloat the installer, make the whole environment much more complex
in order to use jdbc, we would need to start a separate process with JVM, read the data there, then send it between the processes. Slow and complex
given that ODBC and jdbc drivers are often produced by the same company, I personally doubt there's any consistent significant performance advantage of one driver technology over another one; there are usually many other places to look in when chasing performance of data access (I do find some drivers more performant in some cases than others, I do, but I doubt it is caused by the framework)
it might be my ignorance, but I am not aware of any important data source that has a jdbc driver and not an ODBC one
However, I see this request heavily supported, and by people whose opinion I value highly, so, please: Accepting challenges for the above. If there's a business reason, we will find a way.
(Disclaimer: I have used both technologies in the past heavily and I am using both "simultaneously" right now)
I can see this is still not planned, but it really should be considered. There are many databases that do not work well with ODBC connections. We have been advised against using Alteryx to connect to the DB. I would hate to see this be a trend.
do you , please, have some examples for the "There are many databases that do not work well with ODBC connections." and specifically "we have been advised against using Alteryx to connect to the DB"? What is the trouble that you (or the db admin) is experiencing?
Hello @born, current issue trying to connect to SolidDB over ODBC. JDBC works in SQuirrel and DBeaver for connecting. Alteryx and SSIS connects successfully over ODBC. However, When querying, the results from a Select * on any given table will return only one row in Alteryx, whereas all others receive the full output. SolidDB drivers come with ANSI/Unicode versions for both and 32/64 bit. I've tried every combination to get multiple rows. Therefore, potential use case here.
Always best to have options imo. As a workaround to your base installation, why not include Jython library in existing python installation? Create a python input connector to serve as passthrough for connecting over jdbc. Likely wouldn't be as robust as existing input connector, but functional.