commons-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Matt Benson <gudnabr...@gmail.com>
Subject Re: [csv] Object Mapping Proposal
Date Sat, 29 Nov 2014 19:13:52 GMT
I thought the original proposal was more akin to Java beans property
editors, i.e. simple parsing of objects from a string representation. There
are any number of libraries that support this kind of basic
string-to-object conversion. Actually flatfile has some support for this in
its built in Morph [1] integration as well. I also intend, in the future,
to support flatfile in my own Therian [2] object transformation library
(the logistics of this are complicated until/unless flatfile can graduate
from the sandbox), but FWIW I have a preference for pulling CSV through
conversion instead of the other way around. If anything of this nature
should be added to CSV I would recommend considering property editors,
keeping it quite simple to maintain the component's focus.

Matt

[1] http://morph.sourceforge.net/
[2] http://mbenson.github.io/therian/
 On Nov 29, 2014 6:38 AM, "Gary Gregory" <garydgregory@gmail.com> wrote:

> Hibernate is all about mappings. Yes,  it sits on top of JDBC but that
> just means you need a CSV or perhaps Excel driver. There is a lot more
> value in creating/reusing a simple CSV JDBC driver than creating another
> custom mapping/binding framework which duplicates some features of a JDBC
> driver and Hibernate...
>
> My 2c,
> Gary
>
> <div>-------- Original message --------</div><div>From: Benedikt Ritter
<
> britter@apache.org> </div><div>Date:11/29/2014  05:34  (GMT-05:00)
> </div><div>To: Commons Developers List <dev@commons.apache.org>
> </div><div>Cc:  </div><div>Subject: Re: [csv] Object Mapping
Proposal
> </div><div>
> </div>2014-11-28 19:41 GMT+01:00 Gary Gregory <garydgregory@gmail.com>:
>
> > This is probably out of scope for a light weight component like Commons
> > CSV. You could use Hibernate for all your mapping needs.
> >
>
> I don't understand what hibernate has to do with this, since it is an
> object-relational mapping framework and Frank's proposal is about mapping
> between CSV data and pojos... We talked about a mapping functionality for
> csv before. I agree that it is out of scope of what we have currently. but
> I could imagine to transform csv to a multiproject, that has a core
> component and a mapping component (and probably more)
>
> Frank, the mailing list server will strip any attachments. Can you create a
> github repository, so we can discuss your proposal in more detail?
>
> TIA!
> Benedikt
>
>
> >
> > Gary
> >
> > <div>-------- Original message --------</div><div>From: "Ulbricht,
> Frank" <
> > f.ulbricht@qualitype.de> </div><div>Date:11/28/2014  03:13  (GMT-05:00)
> > </div><div>To: dev@commons.apache.org </div><div>Cc:
> </div><div>Subject:
> > [csv] Object Mapping Proposal </div><div>
> > </div>Hello there,
> >
> > about 15 years ago I started to write a CSV library. Our company is using
> > it for a long time now. Over the time a lot of interesting features were
> > added. Now I have decided to replace it with the commons-csv. This API
> > looks very good and it provides some low-level features our library is
> > missing (e.g. multi-line records with escaping).
> >
> > Nevertheless, we have a lot of high-level features for easily mapping
> > between objects and the csv files. I am planning to migrate those to use
> > the commons-csv for the I/O work. And I want to use this chance to
> redesign
> > our APIs.
> >
> > Now a thought crossed my mind, how about contributing this to the
> > commons-csv? In order to get an idea want I am planning to do I have
> > attached a sample project to this mail. It is just an idea, far from
> being
> > perfect. It shall demonstrate how it may look like one day.
> >
> > Sure, in the moment this sample uses a lot of pretty cool Java 8 stuff,
> > but there are ways to make it available to earlier Java versions too.
> >
> > Please have a look at the class “CSVObjectParserTest” it and tell me what
> > you think.
> >
> > Thank you,
> > Frank.
> >
> > P.S. Looks like my previous mail was ignored because the subscription
> > process was not yet finished. If not, please ignore the duplicate mail.
> >
>
>
>
> --
> http://people.apache.org/~britter/
> http://www.systemoutprintln.de/
> http://twitter.com/BenediktRitter
> http://github.com/britter
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message