So I have to write some data files that use EBCDIC and I am excited!!! I have never actually used EBCDIC it is just one of these oddities that I just have never had a chance to use (since I don't do much mainframe work).

Anyway... My first though twas a hash table, but that turns out to be kind of ugly:
And I just see several things that are going to bug me (line byte to Byte and back again). However at this moment using a hash table seems about the best solution there is... I know there are some formulas for converting characters and digits -- but I don't think those work well for symbols.

Anyway, I was wondering if anyone had any advince on the matter.

I am planning on making a stream that I can put data in one and and get it back out the other end converted. This is so I can plug this right into the file io and hopefully write the data transparently converting form ASCII to EBCDIC.

I will make this a snippet once it is done.

This post has been edited by NickDMax: 3 Sep, 2008 - 01:26 PM

Whatever turns you on


Probably that it is still used in the States... all othe languages have accented characters so I guess you will not found EBCDIC used anywhere else... it likes the old AM/PM concept that you will never find on any airline/train/invoice ticket anywhere else in the world



Anyway... My first though twas a hash table, but that turns out to be kind of ugly:


Forget hastable for that, you are using a Cadillac to deliver a pizza
Just build an Array of equivalent characters indexed by the ASCII code will be a lot more efficient
At the limit, if you want to code, write an algorith to build the array that you will run only once at the object creation

This post has been edited by pbl: 3 Sep, 2008 - 03:57 PM

Its a Canadian organization that I am using this with. Actually you will find that Mainframes tend to use this coding a good deal. ("legacy systems" is the dysphemism these days... even though that can refer to any thing from 5 to 100 years old). So lets not jump on this, "Oh, those goofy Americans and their ameri-centric character coding systems!" boat.

Not to mention that EBCDIC is just as "code page" friendly as ASCII (well... maybe not "just as" since the characters are scrambled about). There is even a 2 character wide EBCDIC coding that looks like a Unicode-EBCDIC bastard child.

Anyway! Thanks for the Array idea! This is why I don't like working in Java, I always seem to do things the hard way, just to learn later that there was much easier way to do things.

This post has been edited by NickDMax: 3 Sep, 2008 - 06:03 PM

The String class (and it's .getBytes(encoding) method) of most JDKs support it out of the box, just be careful, because the naming is not really standard according to some. An example

Yea I figured as much! That was why I asked here.

Trust me to go about it the long way again! Thanks