Decoding EBCDIC array to string

I have a file format that returns bytes in EBCDIC format:

3200-element Array{UInt8,1}:
 0xc3
 0x40
 0xf1
 0x40
 0xc3
 0xd3
 0xc9
 0xc5
 0xd5
 0xe3
 0x7a
 0x40
 0xc2
    ⋮
 0x40
 0x40
 0x40
 0x40
 0x40
 0x40
 0x40
 0x40
 0x40
 0x40
 0x40
 0x40

When I copy the bytes before … into an online converter with EBCDIC 0037 formatting, I get the correct value:

Bytes: x" 0XC3 0X40 0XF1 0X40 0XC3 0XD3 0XC9 0XC5 0XD5 0XE3 0X7A 0X40 0XC2 "
ASCII String: Ã@ñ@ÃÓÉÅÕãz@Â
EBCDIC String: C 1 CLIENT: B

I’m trying to do the same within Julia using the StringEncodings.jl package, but getting an error:

decode(ebcdic_array, "EBCDIC-US")

Conversion from EBCDIC-US to UTF-8 not supported by iconv implementation, check that specified encodings are correct

Stacktrace:
 [1] iconv_open(::String, ::String) at /home/rzwitch/.julia/packages/StringEncodings/ICuGB/src/StringEncodings.jl:88
 [2] StringDecoder(::Base.GenericIOBuffer{Array{UInt8,1}}, ::Encoding{Symbol("EBCDIC-US")}, ::Encoding{Symbol("UTF-8")}) at /home/rzwitch/.julia/packages/StringEncodings/ICuGB/src/StringEncodings.jl:286
 [3] decode(::Type{String}, ::Array{UInt8,1}, ::Encoding{Symbol("EBCDIC-US")}) at /home/rzwitch/.julia/packages/StringEncodings/ICuGB/src/StringEncodings.jl:474
 [4] decode(::Array{UInt8,1}, ::String) at /home/rzwitch/.julia/packages/StringEncodings/ICuGB/src/StringEncodings.jl:483
 [5] top-level scope at In[70]:100 

Am I missing an intermediate step?

EDIT: Julia 1.1, StringEncodings 0.3.1

Does StringEncodings.jl support EBCDIC?

"EBCIDIC-US" in encodings() # false

findfirst(x->startswith(x,"EBCDIC"),encodings()) # nothing

This is a very straightforward conversion:

# from https://github.com/Intermernet/ebcdic/blob/master/charmaps.go
const ebcdic_table = Char.([
    0x00, 0x01, 0x02, 0x03, 0x9C, 0x09, 0x86, 0x7F, 0x97, 0x8D, 0x8E, 0x0B, 0x0C, 0x0D, 0x0E, 0x0F, # 0x00..0x0F
    0x10, 0x11, 0x12, 0x13, 0x9D, 0x85, 0x08, 0x87, 0x18, 0x19, 0x92, 0x8F, 0x1C, 0x1D, 0x1E, 0x1F, # 0x10..0x1F
    0x80, 0x81, 0x82, 0x83, 0x84, 0x0A, 0x17, 0x1B, 0x88, 0x89, 0x8A, 0x8B, 0x8C, 0x05, 0x06, 0x07, # 0x20..0x2F
    0x90, 0x91, 0x16, 0x93, 0x94, 0x95, 0x96, 0x04, 0x98, 0x99, 0x9A, 0x9B, 0x14, 0x15, 0x9E, 0x1A, # 0x30..0x3F
    0x20, 0xA0, 0xE2, 0xE4, 0xE0, 0xE1, 0xE3, 0xE5, 0xE7, 0xF1, 0xA2, 0x2E, 0x3C, 0x28, 0x2B, 0x7C, # 0x40..0x4F
    0x26, 0xE9, 0xEA, 0xEB, 0xE8, 0xED, 0xEE, 0xEF, 0xEC, 0xDF, 0x21, 0x24, 0x2A, 0x29, 0x3B, 0xAC, # 0x50..0x5F
    0x2D, 0x2F, 0xC2, 0xC4, 0xC0, 0xC1, 0xC3, 0xC5, 0xC7, 0xD1, 0xA6, 0x2C, 0x25, 0x5F, 0x3E, 0x3F, # 0x60..0x6F
    0xF8, 0xC9, 0xCA, 0xCB, 0xC8, 0xCD, 0xCE, 0xCF, 0xCC, 0x60, 0x3A, 0x23, 0x40, 0x27, 0x3D, 0x22, # 0x70..0x7F
    0xD8, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0xAB, 0xBB, 0xF0, 0xFD, 0xFE, 0xB1, # 0x80..0x8F
    0xB0, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, 0x70, 0x71, 0x72, 0xAA, 0xBA, 0xE6, 0xB8, 0xC6, 0xA4, # 0x90..0x9F
    0xB5, 0x7E, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, 0x79, 0x7A, 0xA1, 0xBF, 0xD0, 0xDD, 0xDE, 0xAE, # 0xA0..0xAF
    0x5E, 0xA3, 0xA5, 0xB7, 0xA9, 0xA7, 0xB6, 0xBC, 0xBD, 0xBE, 0x5B, 0x5D, 0xAF, 0xA8, 0xB4, 0xD7, # 0xB0..0xBF
    0x7B, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, 0x48, 0x49, 0xAD, 0xF4, 0xF6, 0xF2, 0xF3, 0xF5, # 0xC0..0xCF
    0x7D, 0x4A, 0x4B, 0x4C, 0x4D, 0x4E, 0x4F, 0x50, 0x51, 0x52, 0xB9, 0xFB, 0xFC, 0xF9, 0xFA, 0xFF, # 0xD0..0xDF
    0x5C, 0xF7, 0x53, 0x54, 0x55, 0x56, 0x57, 0x58, 0x59, 0x5A, 0xB2, 0xD4, 0xD6, 0xD2, 0xD3, 0xD5, # 0xE0..0xEF
    0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, 0x38, 0x39, 0xB3, 0xDB, 0xDC, 0xD9, 0xDA, 0x9F
])

function ebcdic_to_string(data::AbstractVector{UInt8})
    buf = IOBuffer()
    for byte in data
        print(buf, ebcdic_table[byte+1])
    end
    String(take!(buf))
end

In particular, the above code implements conversion from EBCDIC 037 via this data table. Change ebcdic_table[0x9f + 1] = '€' in order to use EBCDIC 1140 instead.

(There even weirder variants of EBCDIC, e.g. with non-Unicode control characters, that aren’t handled by the above code.)

1 Like

Good point, I looked here and was trying various encodings from the list:

Thanks Steven, I’ll try this out! I suspect this probably should be sufficient for my needs, I wasn’t expecting any wild characters

when I run iconv -l from the command line, I do see “EBCDICUS” listed (no dash). I have version 2.29 of iconv on my machine. There also seems to be a “EBCDIC-CP-US” not sure how that differs.

1 Like

Thanks for checking. The more I read, I think the issue is not that the encoding doesn’t exist, but that iconv is complaining that it won’t convert to UTF-8 like Julia wants.

When I run iconv -l built for the package dependencies, I’m not seeing any EBCDIC encoding support, while my systems library does. So is a build option missing for the binary dependencies?

Good point, I was running iconv -l at the system level; it’s not the same list as ~/.julia/packages/StringEncodings/ICuGB/deps/usr/bin/iconv -l