Python Job Market

Ken kenfar42 at yahoo.com
Sun Jan 4 01:14:43 CET 2004


Samuel Walters <swalters_usenet at yahoo.com> wrote in message news:<pan.2004.01.01.02.57.59.651834 at yahoo.com>...

> You know.  I never thought about this before.  Maybe I could find a niche
> reformatting old cobol records and systems.  Is there any demand for that
> sort of thing? 

Sure - though conversion of batch mainframe applications involves
considerable more complexity than just being able to read COBOL
copybooks.

Perhaps there's more work in data conversion. I've been using python
successfully on data warehouse ETL (extract, transform, load) projects
for a couple of years now.  Best practices are to use a commercial
product for this - but you can easily get hung up in procurement,
training, feasibility studies, approvals, etc - that many projects end
up being implemented in perl instead.  These applications tend to be
maintenance-intensive, are often developed iteratively, have a heavy
focus on data quality, and should be accessible to a wide range of IT
personnel - python is a natural fit.

Anyhow, my current etl project is about 5% korn shell and 95% python
and is processing about 10 million rows/day - and is expected to have
to scale to 300 million rows/day by the end of 2004.  This load will
require some optimizations but can be achieved with the hardware at
hand based on current performance figures.  Note that python wasn't on
the list of approved solutions in this organization, but I was allowed
to use it - since I was doing much of the work in my own time - and
presented the option of an eventual path to Java (via jython).  Now
most of the team is firmly behind it - they've seen how low the
learning curve has been, and how fast it has been to develop with.

ken



More information about the Python-list mailing list