Re[2]: A question about NLS

Archive of postings to DataStageUsers@Oliver.com. This forum intended only as a reference and cannot be posted to.

Moderators: chulett, rschirm

Locked
admin
Posts: 8720
Joined: Sun Jan 12, 2003 11:26 pm

Re[2]: A question about NLS

Post by admin »

At 05:18 PM 1/31/01 +0300, Marat S. Salimov wrote:

Marat,

Unfortunately, Im a long-time VMark/Ardent employee who joined Informix because of the merger last year. Because of this, I dont know a great deal about the GLS implementation.

However, the little I do know leads me to believe you will be pleasantly surprised by the NLS implementation within the DS Server. It is as full-featured an implementation as Ive seen.

One of the great features of NLS is its customization capabilities. Additionally, it allows for multiple levels of control. You can control your NLS settings on a system-wide basis, on an account basis, on an application basis, and even employing different settings within an application.

Now, your questions below lead me to believe that you are interested both
in character set mapping as well as locale support. Let me briefly describe
the two:

- Character Set mapping is just that. The ability to establish
the interpretation of any data stream in use. You use it to
correctly understand any given input and correctly transmit that
data in the correct output format. You can define mappings to
occur at any IO boundary. From the DS Engine perspective, that
includes any client interface, any file/table, any terminal, or
any other device. You can do this globally, per account, or
even individually per run-time instance.

- Locale support is the ability to deal with the DTNC characteristics
of any given country/locale (or as I call it, DTNMCS). Thats
basically:
Dates, Times, Numerics, Monetary, CharacterTypes, and Sorting.


Our NLS support is based on the Posix standard and much like the Char.
set mapping, allows for full customization. In fact, the
customization
is not only allowed, but recommended with locale support.
Whereas character set mapping deals with HOW to interpret characters,
Locales determine WHAT to do with data. Specifically, things
such as:

Dates/Times - how to display the date (is it 01/31/2001, or
31/01/2001, etc...), what to use for the names of the
days/months, how to define am/pm, the era,
who to display the time, etc....
Numeric/Monetary - what the thousands/decimal separators are,
what digits to use for the numerics, the local and
international currency symbols, the positive/negative
sign, and the format of all of the above
Character Types - what the Upper/Lower case versions of a
character are, what are the alphabetic, numeric,
printable, etc.. characters
Sorting - exactly how to sort data within any given language.
whether to care about upper/lower case
differences, how
accented characters sort, how to treat combined
characters
(the german ss for example), how to sort
characters from
different character sets, etc...


NLS support allows you to establish a locale that defines each of this
categories.
More so, it allows you to easily create new locales to satisfy your specific requirements. Currently, we ship with roughly 70 pre-defined locales, based on the Posix standards. For example, we have a Russian locale. However, if
you need
to define a difference in the Russian locale for a particular region, you
can easily
do this as well.

Anyway, hope that long-winded explanation hasnt bored everyone to
tears. I could
go on for days, trust me :-)

As to your other questions:

a) DataStage using IDS repository - Dont know...
b) What features do you lose by not turning on NLS. You lose all of
the above, and more importantly, you may lose the ability to
correctly interpret certain characters, depending on the
map itself.

Again, hope this helps.

Dave




>Hi David,
>
>Thank you for the answer. Is DS (UV) able to operate like IDS with its
>GLS? It would be great!!! But I think it is not. BTW Are there any
>plans about using DataStage somehow wiht the repository on IDS?
>
>Anyway If Ive not turned on NLS and my data sources are NLS-data (by
>ex. CP866, PC1251 etc.) what features Im loosing ? Is it only sorting?
>
>Why these questions youll say? In the case of some "big work" overhead
>in one innecessary byte is significant. Best regards,
> Marat mailto:maratkotik@mail.ru

========================================================================
David T. Meeks || "All my life Im taken by surprise
Development Engineer, DataStage || Im someones waste of time
Ascential Software || Now I walk a balanced line
dave.meeks@ascentialsoftware.com || and step into tomorrow" - IQ
========================================================================
Locked