#Universal database definition how to#
#Universal database definition code#

If I do this right, these are the problems that should be solved by this interface.Adjective Fall temperatures have increased by 2.7 ☏ on average across the contiguous U.S.

#Universal database definition code#

Additionally, if you have a table that gets moved to a different source on a different database (this is something that has actually happened to me a number of times), you will have to actually go in and change your source code to accommodate this.

#Universal database definition how to#

This becomes a problem if you have a dozen different tables living in 3 or 4 different types of databases you will probably have to devote some time to figuring out how to do all the queries. So, you may ask, why bother with this at all of I’m just saying “I’m not going to do any work, DataStreams does it”? Well, as things are right now there is no guarantee of how you get to the point of actually having a Data.Source or Data.Sink. I’ll look into it more and try to make sure I don’t get in its way in the next iteration. Thanks for pointing that out, that is exactly the sort of thing which I would like for it to be possible to extend this to support, even if nothing is written to do so initially. My own situation has improved somewhat and increasingly I am only having to use Postgres, so the usefulness of this for my own personal case has diminished somewhat.Īnd I think good DBAPI has to define also optional specifications to enhance future compatibility (why not to borrow for example two phase commit functionality?). On the other thread there seemed to be a lot of interest in this, but I want to see if it is still there. This is a little iffy because it would involve conditional loading or code-generation.Īny feedback is welcome. Somewhat more ambitious, automatically select a back-end based on configuration toml.I think I’d like to make a standard toml format so that people could just stick a toml in their project with the appropriate connecton strings and instantly have an easy interface for their various databases.Possibly clean up or standardize object constructors so that users of the interface won’t have to change constructor code when swapping out back-ends (this is already possible, but I have to make some choices).Uploading to database (using DataStreams sink objects).(I will contact the maintainers of the JuliaDatabases repositories to see if they are willing to let me use those, I am already an admin on JDBC.) I will have use of this for JDBC and Postgres, so I will commit to JDBC.jl and create a PostgreSQL.jl which adds this simple interface over LibPQ.jl. This is of course very close to the existing database interfaces, so in most cases it will take very little work to get them to implement this. # or you can skip most of the above steps and go straight toĭf = load(DataFrame, cursor(cnxn), 'select * from airlines') # load into DataFrame or any other DataStreams sink Here is an example query using JDBC: # create connectionĬnxn = Connection('jdbc:derby:jar:(toursdb.jar)toursdb') (I’m probably still missing some important bits, but a large class of transactions should be possible with what is already here.) You can find the interface here and an example implementation in JDBC here. Note that I have no intention of having any sort of SQL-like domain-specific query language. Execute a query from an AbstractString containing SQL code.Connect to the database using only a connection AbstractString.My goal was to provide an interface which allows you to change database back-ends by only changing your connection string and a single using or import statement. It is so minimal, in fact, that by design you should not even have to import this package to implement it. Hello all, I have written an extremely minimal universal database API in the style of PEP249.