Hi all,
I currently have a very simple, fairly small (25,000 records), one table
database in MS SQL Server that is accessed for read only & read / write
(dependant on user interface) via ASP on IIS (running on NT4 SP6).
I am considering moving the from ASP to PERL CGI for accessing & rendering
the output and am also considering the pros / cons of abadoning RDMS in
favor of flat file(s) and was wondering about comparative performance of
flat file options compared to an RDMS for such a simple db. Unfortunately
NT4 & IIS are a must - no amount of open source / linux / apache advocacy
has changed this :(
Each record has a primary key composed of three of the fields (these 3 are
fixed length) and ach record is approximately 2Kb in size (most of which is
in a narrative field). The data is accessed in two ways:
1. Read only list of records - selected by setting optional criteria of
a couple of fields (one of which is one of the fields making up the rimary
key, the others are indexed non-key fields)
2. Read / Write access to a single record, selected by entering the
primary key
Options (as I see them):
1. Status quo - using RDMS
2. Large delimited flat file containg all records (are there concurrency
issues with locking, updating, etc)
3. Large xml file containing all records (imagine this would be fairly
inefficient in addition to 2's issues)
4. Large numer of small text files, each containg details of a single record
(files named after primary key of record)
Is having a seperate index file for searching by non-key fields practical
for options 2 & 4 ... I've already discounted 3 as being a serious
contender.
Your thoughts & experiences are warmly welcomed!
Thanks,
Andy
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
<http://learn.perl.org/> <http://learn.perl.org/first-response>
