EBCDIC <--> ASCII

[martinjamesevans]

I'm having a problem trying to use the codecs package to aid me in
converting some bytes from EBCDIC into ASCII.

I have some 8bit text that is in mixed format. I extract the bytes
that are coded for EBCDIC and would like to display them correctly.
The bytes that are EBCDIC could values 0-255, I'm only really
interested in the printable portions and could say leave the rest as
dots.

I've tried starting with something like this, but I assume it is
expecting the source to be in unicode already?

e.g. (pretend the second half are EBCDIC characters)

sAll = "This bit is ASCII, <this bit ebcdic>"
sSource = sAll[19:]

sEBCDIC = unicode(sSource, 'cp500', 'ignore')
sASCII = sEBCDIC.encode('ascii')

Obviously I could just knock up a 255 character lookup table and do it
myself, I was just trying to be a little more Pythonic and use that
built in table.

Thanks,

Martin

 

[Ulrich Eckhardt]

I've tried starting with something like this, but I assume it is
expecting the source to be in unicode already?

e.g. (pretend the second half are EBCDIC characters)

sAll = "This bit is ASCII, <this bit ebcdic>"

Why pretend? You can use this:

"abcde\x81\x82\x83\x84"

sSource = sAll[19:]

sEBCDIC = unicode(sSource, 'cp500', 'ignore')

If you mean this sSource, then no. sSource is treated as byte string here
which is converted to Unicode using 'cp500' as encoding. Note that in
interactive mode, 'print x' will actually convert the string according to
stdout's current encoding (typically ASCII or - I think - Latin 1).

s1 = u'abcde'
s2 = s1.encode('cp500')
s3 = s1.encode('ascii')
s4 = unicode( s2, 'cp500')
s5 = unicode( s3, 'ascii')

Uli

 

[Michael Ströder]

I'm having a problem trying to use the codecs package to aid me in
converting some bytes from EBCDIC into ASCII.

Which EBCDIC variant?

sEBCDIC = unicode(sSource, 'cp500', 'ignore')

Are you sure CP500 is the EBCDIC variant for the language you want?

http://www.ietf.org/rfc/rfc1345.txt lists it as:

&charset IBM500
&rem source: IBM NLS RM Vol2 SE09-8002-01, March 1990
&alias CP500
&alias ebcdic-cp-be
&alias ebcdic-cp-ch

Obviously I could just knock up a 255 character lookup table and do it
myself, I was just trying to be a little more Pythonic and use that
built in table.

It's pythonic to implement a Unicode codec for unknown character tables.
I've put these two on my web site:

http://www.stroeder.com/pylib/encodings/ebcdicatde.py
http://www.stroeder.com/pylib/encodings/cp273.py (needs ebcdicatde)

Ciao, Michael.

 

[Michael Ströder]

Thanks for the tables, ebcdicatde.py does look more suitable.

My problem appears to be that my source is a byte string. In a
nutshell I need "\x81\x82\x83\xf1\xf2\xf3" to become "abc123" in a
byte string.

Python 2.5.2 (r252:60911, Aug 1 2008, 00:43:38)
Ciao, Michael.

 

[martinjamesevans]

Python 2.5.2 (r252:60911, Aug  1 2008, 00:43:38)
 >>> import ebcdicatde
 >>> "\x81\x82\x83\xf1\xf2\xf3".decode('ebcdic-at-de').encode('ascii')
'abc123'
 >>>

Ciao, Michael.- Hide quoted text -

- Show quoted text -

Many thanks for all your posts!
Just what I needed.