Re: Output to text files or a database?
- From: Lew Pitcher <Lew.Pitcher@xxxxxxxxxxxxxxxx>
- Date: Wed, 01 Mar 2006 12:19:05 -0500
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
kd wrote:
I was reading The Art of Unix Programming yesterday, and was inspired
to rework and split up an large monolithic application of mine.
This software has to do with the retrieval and processing of travel
information (flights, hotels, etc) from different sources. Each source
has travel data formatted differently.
With plain text files being the universally loved file format, I
planned on writing small apps that would pull down data from these
different sources, and write them to text files in a universal format.
Another small app would read these text files, generate sql insert
statements from the data, and dump all the data into a relational
database.
I proposed this to a coworker, and he asked why the text files were
even necessary -- why not just immediately generate sql statements from
the different formats, and use the tables in the DB as the universal
format? Textual representations of that data could just be generated
from the database, after all.
Is he right? Is there any reason to use text files as an intermediate
format?
I don't know. But I'd have some questions, like
- - what do you do if the database server isn't available when the app
needs to store it's data?
- - what do you do if the database server isn't available when the app
that needs the data requests it?
- - does the data need to be preprocessed, hand edited, or otherwise
subjected to non-db processing?
- - database apps are largish. Are you getting into the same sort of
maintenance issue that you are moving out of, by building one larger
app rather than two smaller apps?
- - is there any reason to want to perform ad-hoc analysis on the text
files? can you do the same against the database? How easily?
- - is there any reason to want to perform alternate processing on the
data (does it or can it serve any other purpose other than what you
intend)? can you do this processing on text files? how easily, as
compared to against a database?
- - what about recoverability? Auditability? Are these important?
My gut is telling me to keep them, but my gut isn't good at
convincing people. :) What am I missing?
- --
Lew Pitcher, IT Specialist, Corporate Technology Solutions,
Enterprise Technology Solutions, TD Bank Financial Group
(Opinions expressed here are my own, not my employer's)
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (MingW32)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
iD8DBQFEBdeJagVFX4UWr64RAkx+AKDi3OvUbWHOrnOAGPjoE2N9glC77QCgzp5y
ww+bkaEz2x2/bq7CqRfiog0=
=muEF
-----END PGP SIGNATURE-----
.
- References:
- Output to text files or a database?
- From: kd
- Output to text files or a database?
- Prev by Date: Output to text files or a database?
- Next by Date: Re: Output to text files or a database?
- Previous by thread: Output to text files or a database?
- Next by thread: Re: Output to text files or a database?
- Index(es):
Relevant Pages
|