Mailing List Archive

OT: Tie::DBI / DBIxRecordset in embperl
hi,

maybe its dead off-topic, but as it's 'offered' as an embperl
complement, I'd like to know about your experiences in using
DBIx::Recordset . I'm currently using a DBI with DBD::mysql and Tie::DBI
for 'quicky' access and pure DBI for real work.

does it offer any advantage? I read the manpage and it looked like a
Tie::DBI twin, with maybe some magic ties with embperl (that little
magic that saves the day).



martin
Re: OT: Recordset [ In reply to ]
Martin Langhoff wrote:
>
> maybe its dead off-topic, but as it's 'offered' as an embperl
> complement, I'd like to know about your experiences in using
> DBIx::Recordset.

Gerald, I wanted to ask the same: is Recordset on topic here? I believe
it's OK since Recordset is designed for Embperl and used mostly together
and anyway, you don't want to read two lists if can avoid it, would you?
What do you think and what other people on this list think?

> I'm currently using a DBI with DBD::mysql and Tie::DBI
> for 'quicky' access and pure DBI for real work.
> does it offer any advantage? I read the manpage and it looked like a
> Tie::DBI twin, with maybe some magic ties with embperl (that little
> magic that saves the day).

In my experience Recordset is extremly cool but still quite young. That
means it's mostly working but it lacks the polish of the finished
product: while I don't think I have found any real bug in Embperl ever
(since 1.0), with Recordset it's still quite often. I started with
Recordset to get things done fast and while writing _is_ much faster
than plain DBI, debugging can be sometime a bit painfull. On the other
hand Recordset has improved rather fast lately and B21 is supposed to be
much better (and B20 already was much better then B19 I started to work
with). Also I hear docs are much better in B21, which is also a big win.

Of four my projects I tried to use Recordset for only one actually is
Recordset based, but it works and going online next week. I'm planning
to use Recordset much more in the future and I don't mind spending some
time working around current Recordset problem (and maybe even finish few
patches I promised Gerald long ago ;-). Hope this helps.

- Robert
RE: OT: Recordset [ In reply to ]
> > maybe its dead off-topic, but as it's 'offered' as an embperl
> > complement, I'd like to know about your experiences in using
> > DBIx::Recordset.
>
> Gerald, I wanted to ask the same: is Recordset on topic here? I believe
> it's OK since Recordset is designed for Embperl and used mostly together
> and anyway, you don't want to read two lists if can avoid it, would you?
> What do you think and what other people on this list think?
>

For me it's on topic here. I don't think it will increase the traffic too
much, but if the questions are more DBI related then Web related, it maybe
better to post them to the DBI mailing list. It's up to you were you think
it fits better.

I am subscribed to both (and much more mailing lists) and also when I am too
busy to review the subject lines my Mailprogramm sorts anything out which
contains Embperl or Recordset. So I will get them anyway.

Gerald
RE: Tie::DBI / DBIxRecordset in embperl [ In reply to ]
hi,
>
> maybe its dead off-topic, but as it's 'offered' as an embperl
> complement, I'd like to know about your experiences in using
> DBIx::Recordset . I'm currently using a DBI with DBD::mysql and Tie::DBI
> for 'quicky' access and pure DBI for real work.
>
> does it offer any advantage? I read the manpage and it looked like a
> Tie::DBI twin, with maybe some magic ties with embperl (that little
> magic that saves the day).
>

Then you should read it a second time! DBIx::Recordset offers much more then
simply tieing a database table to hash or array. It's a database abtraction
layer above DBI and it simplify database access.

Read http://perl.apache.org/embperl/Intrors.pod.cont.html (or
http://www.heise.de/ix/artikel/1999/09/137/ for the german version) for an
Introduction. I append you the list of the features of DBIx::Recordset, to
get an first impression.

Gerald


The main features of DBIx::Recordset are:

- it has a compact interface, normaly only one function call is necessary
for setup and data retrival/inseration/deletion

- it takes care about type conversion and quoting

- it is able to access/modify tables via arrays and hashs

- it can automaticly create sub-objects for tables which are logical linked
together

- it can automatily create joins based on logical links

- it has input/output filters on a per field/per type basis

- it can create WHERE expression from a hash, which is especially usefull in
a cgi environement, where you can simply pass all paramters posted to your
cgi
script to DBIx::Recordset and DBIx::Recordset creates an corresponding
SELECT.

- it can create previous/next buttons for html output

- it works together with HTML::Embperl for easily genration of HTML output

- it has an own database abtraction class DBIx::Compat which gives all the
necessary information, so that DBIx::Recordset is able to work with
different database systems

- The new class DBIx::Database is able to retrieve and store meta infomation
of the database in a centralised location, which can be used for later
setup. This is also usefull when running under mod_perl, because you can
do
all the setup and configuration work at webserver startup time, speeding
up
your scripts when a actual request is processed.


-------------------------------------------------------------
Gerald Richter ecos electronic communication services gmbh
Internetconnect * Webserver/-design/-datenbanken * Consulting

Post: Tulpenstrasse 5 D-55276 Dienheim b. Mainz
E-Mail: richter@ecos.de Voice: +49 6133 925151
WWW: http://www.ecos.de Fax: +49 6133 925152
-------------------------------------------------------------