It might be simpler to either “brew install mdbtools” or “apt-get install mdbtools” and use mdb-export [mdb file] [table name] > [table_name].csv
Will > On Jan 19, 2016, at 9:19 PM, P Jakobsen <pjakob...@gmail.com> wrote: > > Hi William, > > >> On Jan 19, 2016, at 9:48 PM, William Witt <will...@witt-family.net> wrote: >> >> I think the answer is no. It's theoretically possible to build a storage >> plugin to handle it using a library like jackcess, but I don't see it as a >> common need. Can you flesh out your use case a little? > > I’m a casual Drill user. The main benefit is that I can dump files that I > download from various open data sites, and just dump them to my drive, write > some queries, and create parquet file. These are mostly files on > international trade data, which tends to come in thousands of cvs files that > unzip into hierarchies. Now I need to join these with custom tariff data, > which sadly is only available as an Access .mdb file. Since this file is > updated on a regular basis, I was hoping to just write a python script that > grabs the file from the website, and dumps into my drill data directory. > > All these files on trade have a common key, which is called an Harmonized > System code. That’s how I join the data across different file formats and > sources. > > I’m an OS X and Linux user, so it would be a hassle for me to learn how to > write some wonky script that coverts .mdb to a sane readable format. > > I hope this all makes sense. > > Peder > >> >> Will >> >> Sent from my iPhone >> >>> On Jan 19, 2016, at 8:36 PM, Peder Jakobsen | gmail <pjakob...@gmail.com> >>> wrote: >>> >>> I have an .mdb (Access file) that ideally I'd like for Drill to see as just >>> another data source. Is this possible? >>> >>> It seems like the MapR driver is for connecting to Drill via ODBC only, >>> which doesn't cover my use case. >>> >>> Thanks :) >
