NLS Setting Issue

[MaheshKumar Sugunaraj]

Hi All,

I have been a special case / issue with respect to the Source Data,

I am currently working on the Baltic's Countries and I have found a strange issue which is as follows:

Č

Ū

1. Could anybody please tell me what kind of NL Setting is suitable to get the data in a Proper way.

2. Could anybody please tell me how to replace Values such as above with an equivalent value as give

Č = C

Ū = U
So on.

The reason we are doing this is because we need to do grouping on a specific column as well.

Any help would be highly appreciated. Many thanks in advance.

Thanks & Regards
Mahesh

[ArndW]

Mahesh,

you do not necessarily need to activate NLS just in order to handle these two characters. They fit into the single-byte space quite well and if you need to do conversions then you could always do a CONVERT("ČŪ","CU",MyString) function on them and their ilk; this would be easier than activating and using NLS. In addition NLS will map these characters and will not perform a conversion for you.

[MaheshKumar Sugunaraj]

Hi,

Thanks for your thoughts, These are a few alphabets which I have given as an example they are lot many.

The NLS Setting we have been currently using is ISO Standard and in some case we have been using PC852 but this will help us to resolve for 80% of the data but the rest 20% are also sensitive and important for analysis, So please give me ur thoughts or Ideas.

Please let me know your thoughs or Ideas regarding the same.

Also Please let me know how to Convert such Characters to a regualer English Alphabet which could be used for Grouping.

Thanks & Regards
Mahesh

[ArndW]

Mahesh,

if you need to collate, then the LOCALE setting is important, you might choose an eastern european locale (Czech) and those character would get collated according to their definition in that locale - chances are that this will be an equivalent to their western counterpart; but you would have to look at the locales to be sure. The conversion might also work if your output NLS setting is to a western-only character set. This again depends upon the character set's definition file (which you can modify if you wish, it is documented in the NLS user's guide). You will know if this doesn't work, as the NLS settings will give you a warning about unmappable characters.

Basically there is no 100% guaranteed method to automatically do all of these conversions for you, especially as you have indicated that you are getting multiple locales (the a-umlaut gets sorted in different locations in different locales, and I think that the nordic transliteration is not "ae" as it is from German).

You might write a mapping function which checks to see if a string is displayable using the ALPHA function (careful, it knows about NLS settings) or the "MCP" conversion code with ICONV/OCONV and then do a big CASE for those characters which you need to convert.

[MaheshKumar Sugunaraj]

Hi,

Thanks again for your thoughts, The problem is the special characters are in the string and this is not a Individual alphabet and I doubt whether I could take a risk of converting sensitive information.

But I just found from Ascential Support that some company's have rasied a E Case ticket for this to be implemented in the DS Server but Luckily this code page is available in the Parallel Extender, But we donot use Parallel Extender.

If you'll have any information of how to go around this please do share with me.

Many thanks in advance.

Regards
Mahesh