(Sorry for the duplicate, Jamie, forgot to forward to
the list!)
--- Jamie <[EMAIL PROTECTED]> wrote:
> ...
> 2) Should i use text files for my data or dive
> straight into something
> like MySQL? I'll propably need the SQL stuff later
> on, but is there
> any general rule as regards to size/ number of
> lines/records?
> ...
I've just gone through a similar decision process. It
depends entirely on your situation. Some questions you
need to consider are:
1) What tools does my platform support?
2) How much data do I have now, and how fast will it
grow?
3) How many users do I envision, both concurrent
(for record locking, etc.) and total (when security
gets more important)?
I would definately recommend some kind of SQL database
solution for most commercial, and certainly
enterprise-wide, applications. I've heard good things
about mySQL, but not used it. At the high end, Oracle
is one of the most solid there is.
I can't quote any exact quantitative recommendations,
but I personally wouldn't consider text files for
anything bigger than a dozen files of maybe a few
hundred records in each. For one thing, in any system
of any size, ease and efficiency of record selection
and sorting becomes very important, and SQL makes
these things very easy. There are SQL drivers for text
files, but I would still be concerned with how
efficient they could be for large amounts of data. At
some point, you've got to have indexes.
On the other extreme, there are cases where a SQL
database may be overkill. One of my personal projects
is a very basic content management system for
non-profits and small businesses, tailored to outfits
with modest sites virtual-hosted by third-party ISPs
and with up to 1/2 dozen content managers, maybe
volunteers. I needed to find a storage mechanism that
would be easily supported by most ISPs. Some IPSs
support mySQL, but not all, and I wanted to be as
compatible as possible. (I selected Perl as a language
due to its almost universal support by ISPs.) I
seriously considered XML storage, but decided that my
own learning curve in using it would be too steep for
the timetable I wanted to meet, and Perl-based XML
tools seem to be in limited supply. So I settled on
text CSV files. For my needs, this seemed to make
sense. The AnyData::CSV (see CPAN) module supports
record selection from CSV's, and I implemented a
multi-level sort just using Perl's sort function. (I
think it also has a DBI interface option, which I have
not investigated.)
So those are some considerations for you. In the end,
it depends on your situation and it's a judgement
call.
- John
__
Do You Yahoo!?
Yahoo! Games - play chess, backgammon, pool and more
http://games.yahoo.com/
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]