Bringing MainFrames back to live with Julia converting ebcdic to text file

[siralf]

Hello Geeks,
Some guidance would enlighten my dark world.
I have ebcdic files coming in and need to covert these to ASCII so I can push/digest these to other systems like Oracle and Kafka for processing...

So, I'm wondering if I can write these functions in Julia. It's said that it's super fast.
I found this closed post. Credit to Rick, he did a great job with these two functions:

https://www.tek-tips.com/viewthread.cfm?qid=52059#post

*** ------------------------------------------------------------------ ***
***
*** These routines convert ASCII (ANSI) and EBCDIC strings.
***
*** asc2ebc(str) : ASCII String to EBCDIC String
***
*** ebc2asc(str) : EBCDIC String to ASCII String
***
*** EXAMPLES:
***
*** ? ebc2asc(asc2ebc('0123456789'))
*** ? ebc2asc(asc2ebc('abcdefghijklmnopqrstuvwxyz'))
*** ? ebc2asc(asc2ebc(UPPER('abcdefghijklmnopqrstuvwxyz')))
*** ? ebc2asc(asc2ebc('!@#$%^&*()_+{}:"<>?|~-=[];,./`'+"'"))
***
*** ------------------------------------------------------------------ ***
FUNCTION asc2ebc && Convert ASCII to EBCDIC
PARAMETER zcasc
PRIVATE lcebc, lnii, lntmp

*** Note: There is no EBCDIC '!' or '^' ***
IF TYPE('gcchrebc') <> "C"
RELEASE gcchrebc && jic
PUBLIC gcchrebc
gcchrebc = ""
* - NUL -> SI
gcchrebc = gcchrebc+chr(000)+chr(001)+chr(002)+chr(003);
+chr(055)+chr(045)+chr(046)+chr(047)
gcchrebc = gcchrebc+chr(022)+chr(005)+chr(021)+chr(011);
+chr(012)+chr(013)+chr(014)+chr(015)
* - DLE -> US
gcchrebc = gcchrebc+chr(016)+chr(017)+chr(018)+chr(019);
+chr(060)+chr(061)+chr(050)+chr(038)
gcchrebc = gcchrebc+chr(024)+chr(025)+chr(063)+chr(033);
+chr(028)+chr(029)+chr(030)+chr(031)
* - <space> -> /
gcchrebc = gcchrebc+chr(064)+chr(255)+chr(127)+chr(123);
+chr(091)+chr(108)+chr(080)+chr(125)
gcchrebc = gcchrebc+chr(077)+chr(093)+chr(092)+chr(078);
+chr(107)+chr(096)+chr(075)+chr(097)
* - 0 -> ?
gcchrebc = gcchrebc+chr(240)+chr(241)+chr(242)+chr(243);
+chr(244)+chr(245)+chr(246)+chr(247)
gcchrebc = gcchrebc+chr(248)+chr(249)+chr(122)+chr(094);
+chr(076)+chr(126)+chr(110)+chr(111)
* - @ -> O
gcchrebc = gcchrebc+chr(124)+chr(193)+chr(194)+chr(195);
+chr(196)+chr(197)+chr(198)+chr(199)
gcchrebc = gcchrebc+chr(200)+chr(201)+chr(209)+chr(210);
+chr(211)+chr(212)+chr(213)+chr(214)
* - P -> _
gcchrebc = gcchrebc+chr(215)+chr(216)+chr(217)+chr(226);
+chr(227)+chr(228)+chr(229)+chr(230)
gcchrebc = gcchrebc+chr(231)+chr(232)+chr(233)+chr(074);
+chr(224)+chr(090)+chr(255)+chr(109)
* - ` -> o
gcchrebc = gcchrebc+chr(121)+chr(129)+chr(130)+chr(131);
+chr(132)+chr(133)+chr(134)+chr(135)
gcchrebc = gcchrebc+chr(136)+chr(137)+chr(145)+chr(146);
+chr(147)+chr(148)+chr(149)+chr(150)
* - p -> DEL
gcchrebc = gcchrebc+chr(151)+chr(152)+chr(153)+chr(162);
+chr(163)+chr(164)+chr(165)+chr(166)
gcchrebc = gcchrebc+chr(167)+chr(168)+chr(169)+chr(192);
+chr(079)+chr(208)+chr(161)+chr(007)
ENDIF
lcebc = ""
FOR lnii = 1 to LEN(zcasc)
lntmp = ASC(SUBSTR(zcasc, lnii, 1))+1 && (+1 from zero to one based)
IF lntmp < 128 && OK to Translate
lcebc = lcebc + SUBSTR(gcchrebc, lntmp, 1)
ELSE && No direct equivalance
lcebc = lcebc + chr(255)
ENDIF
ENDFOR

RETURN lcebc

*** ------------------------------------------------------------------ ***

FUNCTION ebc2asc && Convert EBCDIC to ASCII
PARAMETER zcebc
PRIVATE lcasc, lnii

IF TYPE('gcchrasc') <> "C"
RELEASE gcchrasc && jic
PUBLIC gcchrasc

gcchrasc = ""
* - NUL -> SI
gcchrasc = gcchrasc+chr(000)+chr(001)+chr(002)+chr(003);
+chr(255)+chr(009)+chr(255)+chr(127)
gcchrasc = gcchrasc+chr(255)+chr(255)+chr(255)+chr(011);
+chr(012)+chr(013)+chr(014)+chr(015)
* - DLE -> US
gcchrasc = gcchrasc+chr(016)+chr(017)+chr(018)+chr(019);
+chr(255)+chr(255)+chr(008)+chr(255)
gcchrasc = gcchrasc+chr(024)+chr(025)+chr(255)+chr(255);
+chr(028)+chr(029)+chr(030)+chr(031)
* - LF -> BEL
gcchrasc = gcchrasc+chr(255)+chr(255)+chr(255)+chr(255);
+chr(255)+chr(010)+chr(023)+chr(027)
gcchrasc = gcchrasc+chr(255)+chr(255)+chr(255)+chr(255);
+chr(255)+chr(005)+chr(006)+chr(007)
* - SYN -> SUB
gcchrasc = gcchrasc+chr(255)+chr(255)+chr(022)+chr(255);
+chr(255)+chr(255)+chr(255)+chr(004)
gcchrasc = gcchrasc+chr(255)+chr(255)+chr(255)+chr(255);
+chr(020)+chr(021)+chr(255)+chr(026)
* - <space> -> Vert Bar
gcchrasc = gcchrasc+chr(032)+chr(255)+chr(255)+chr(255);
+chr(255)+chr(255)+chr(255)+chr(255)
gcchrasc = gcchrasc+chr(255)+chr(255)+chr(091)+chr(046);
+chr(060)+chr(040)+chr(043)+chr(124)
* - & -> Not
gcchrasc = gcchrasc+chr(038)+chr(255)+chr(255)+chr(255);
+chr(255)+chr(255)+chr(255)+chr(255)
gcchrasc = gcchrasc+chr(255)+chr(255)+chr(093)+chr(036);
+chr(042)+chr(041)+chr(059)+chr(255)
* - - -> ?
gcchrasc = gcchrasc+chr(045)+chr(047)+chr(255)+chr(255);
+chr(255)+chr(255)+chr(255)+chr(255)
gcchrasc = gcchrasc+chr(255)+chr(255)+chr(255)+chr(044);
+chr(037)+chr(095)+chr(062)+chr(063)
* - ` -> "
gcchrasc = gcchrasc+chr(255)+chr(255)+chr(255)+chr(255);
+chr(255)+chr(255)+chr(255)+chr(255)
gcchrasc = gcchrasc+chr(255)+chr(096)+chr(058)+chr(035);
+chr(064)+chr(039)+chr(061)+chr(034)
* - a -> i
gcchrasc = gcchrasc+chr(255)+chr(097)+chr(098)+chr(099);
+chr(100)+chr(101)+chr(102)+chr(103)
gcchrasc = gcchrasc+chr(104)+chr(105)+chr(255)+chr(255);
+chr(255)+chr(255)+chr(255)+chr(255)
* - j -> r
gcchrasc = gcchrasc+chr(255)+chr(106)+chr(107)+chr(108);
+chr(109)+chr(110)+chr(111)+chr(112)
gcchrasc = gcchrasc+chr(113)+chr(114)+chr(255)+chr(255);
+chr(255)+chr(255)+chr(255)+chr(255)
* - ~ -> z
gcchrasc = gcchrasc+chr(255)+chr(126)+chr(115)+chr(116);
+chr(117)+chr(118)+chr(119)+chr(120)
gcchrasc = gcchrasc+chr(121)+chr(122)+chr(255)+chr(255);
+chr(255)+chr(255)+chr(255)+chr(255)
* - .. ..
gcchrasc = gcchrasc+chr(255)+chr(255)+chr(255)+chr(255);
+chr(255)+chr(255)+chr(255)+chr(255)
gcchrasc = gcchrasc+chr(255)+chr(255)+chr(255)+chr(255);
+chr(255)+chr(255)+chr(255)+chr(255)
* - { -> I
gcchrasc = gcchrasc+chr(123)+chr(065)+chr(066)+chr(067);
+chr(068)+chr(069)+chr(070)+chr(071)
gcchrasc = gcchrasc+chr(072)+chr(073)+chr(255)+chr(255);
+chr(255)+chr(255)+chr(255)+chr(255)
* - } -> R
gcchrasc = gcchrasc+chr(125)+chr(074)+chr(075)+chr(076);
+chr(077)+chr(078)+chr(079)+chr(080)
gcchrasc = gcchrasc+chr(081)+chr(082)+chr(255)+chr(255);
+chr(255)+chr(255)+chr(255)+chr(255)
* - \ -> Z
gcchrasc = gcchrasc+chr(092)+chr(255)+chr(083)+chr(084);
+chr(085)+chr(086)+chr(087)+chr(088)
gcchrasc = gcchrasc+chr(089)+chr(090)+chr(255)+chr(255);
+chr(255)+chr(255)+chr(255)+chr(255)
* - 0 -> 9
gcchrasc = gcchrasc+chr(048)+chr(049)+chr(050)+chr(051);
+chr(052)+chr(053)+chr(054)+chr(055)
gcchrasc = gcchrasc+chr(056)+chr(057)+chr(255)+chr(255);
+chr(255)+chr(255)+chr(255)+chr(255)
ENDIF

lcasc = ""
FOR lnii = 1 to LEN(zcebc)
* && (+1 from zero based to one based)
lcasc = lcasc + SUBSTR(gcchrasc, ASC(SUBSTR(zcebc, lnii, 1))+1, 1)
ENDFOR

RETURN lcasc

*** ------------------------------------------------------------------ ***

 

[Olaf Doschke]

So what's your question?

Use any language you like and know, look up the definition of EBCDIC and go for it:

https://en.wikipedia.org/wiki/EBCDIC

First google result:

https://discourse.julialang.org/t/decoding-ebcdic-array-to-string/26192

You want fast? Either go for assembler or C/C++. Any new language is built on top of these and by definition always a higher-level language with more comfort and less speed And CPUs still run assembler only.
But mapping some byte combinations to others is no big deal.

Just, why reinvent the wheel when other languages support EBCDIC?

https://docs.microsoft.com/en-US/dotnet/api/system.text.encodinginfo?view=netcore-3.1

Bye, Olaf.

Olaf Doschke Software Engineering

https://www.doschke.name

 

[Ljupce Zagorac]

Siraf,

As Olaf mentioned for fast conversion go to C/C++. If an EBCDIC file contains packed fields you'll have to describe each column starting and ending position and, therefore, overall record width.