INPUTCSV example

[Carl Gundel]

Someone emailed me asking about the INPUTCSV command, so I wrote this short example.  I hope it is helpful.


   'write some CSV lines into a file
   open "testfile.csv" for output as #csvfile
   print #csvfile, "1,this,is,a,test"
   print #csvfile, "2,and,a,second,test"
   print #csvfile, "3,and,one,more,test"
   close #csvfile

   'now read the CSV data and print it out
   open "testfile.csv" for input as #readcsv
   while not(eof(#readcsv))
       inputcsv #readcsv, num, a$, b$, c$, d$
       print num
       print a$
       print b$
       print c$
       print d$
       print
   wend
   close #readcsv

   print "Done."
   end


[Carl Gundel]

Oct 22, 2020 13:13:50 GMT -5 Carl Gundel said:

Someone emailed me asking about the INPUTCSV command, so I wrote this short example.  I hope it is helpful.


   'write some CSV lines into a file
   open "testfile.csv" for output as #csvfile
   print #csvfile, "1,this,is,a,test"
   print #csvfile, "2,and,a,second,test"
   print #csvfile, "3,and,one,more,test"
   close #csvfile

   'now read the CSV data and print it out
   open "testfile.csv" for input as #readcsv
   while not(eof(#readcsv))
       inputcsv #readcsv, num, a$, b$, c$, d$
       print num
       print a$
       print b$
       print c$
       print d$
       print
   wend
   close #readcsv

   print "Done."
   end

Here is a link to a video I made about this subject last year.  youtu.be/VDpK3Nh4Sik

[alincon]

I think the example illustrates the need for a 'printcsv' command.

r.m.

[Carl Gundel]

Oct 24, 2020 10:49:59 GMT -5 alincon said:

I think the example illustrates the need for a 'printcsv' command.

r.m.

It's a lot easier to print in CSV than it is to read it, so it's not so compelling.

[xcoder]

Carl, your "dechex$" is not true hexidecimal. I had to add some code the make it "true hexidecimal."
rem Main Program
Print "######################################"
for i = 0 to 255
char$ = char$ + chr$(i)
next i
print char$
print
a$ = text2Hex$(char$)
print a$
print
b$ = hex2Text$(a$)
print b$

print:print
print "End of Code"
End


function text2Hex$(text$)
for i=1 to len(text$)
hex$ = dechex$(asc(mid$(text$, i, 1)))
fullHex$ = fullHex$ + trueHex$(hex$)
next i
text2Hex$ = fullHex$
end function

function trueHex$(a$)
if a$="0" then a$="00"
if a$="1" then a$="01"
if a$="2" then a$="02"
if a$="3" then a$="03"
if a$="4" then a$="04"
if a$="5" then a$="05"
if a$="6" then a$="06"
if a$="7" then a$="07"
if a$="8" then a$="08"
if a$="9" then a$="09"
if a$="A" then a$="0A"
if a$="B" then a$="0B"
if a$="C" then a$="0C"
if a$="D" then a$="0D"
if a$="E" then a$="0E"
if a$="F" then a$="0F"
trueHex$=a$
end function

function hex2Text$(hex$)
for i=1 to len(hex$)step 2
chunk$ = mid$(hex$,i,2)
ascii = hexDec(chunk$)
char$ = char$ + chr$(ascii)
next i
hex2Text$ = char$
end function

[Chris Iverson]

By your logic, all the numbers under "10" in the decimal system by themselves aren't "true numbers".

It's not "2 + 2 = 4", it's "02 + 02 = 04".

It's not "5 + 5 = 10", it's "05 + 05 = 10".

Seems silly, doesn't it?

The thing to keep in mind is that a hexadecimal number is just that: a number. Leading zeroes are meaningless in numbers, and are generally only included for spacing/sorting purposes.

dechex$ works perfectly, and it properly converts any decimal number you give it to the proper corresponding hexadecimal representation.

The problem is you're assigning meaning to the output that isn't there. You're trying to get representations of bytes from dechex$(), NOT representations of numbers. There is an important distinction there; dechex$() ONLY works with numbers, and ONLY returns numbers. If you want those numbers to mean something else, you have to do the conversion yourself.

After all, what number of leading zeroes is "proper"? One zero, to make sure all numbers returned are two-digit byte representations?

Well, what if I don't want bytes? What if I want 32-bit words? "dechex$() doesn't work properly, 254 should be "000000FE", not "FE", it's broken."

Heck, it gets worse than that.

"000000FE" is a 32-bit word, in big-endian order. (See en.wikipedia.org/wiki/Endianness for an overview of what this means.) What if I want it displayed in little-endian order, like x86(and Windows) use internally? Then it's "dechex$ is wrong, it returns "000000FE" for 254 instead of "FE000000". It doesn't format the numbers properly."

It's not supposed to. It's assigning meaning to numbers that isn't there.

You're breaking the bytes of a string up into individual values, converting the numbers into a different format, and then reassembling those values. It's up to you to make sure the meaning stays consistent when doing that, because dechex$() only works with numbers.

Now, if LB had some sort of built-in function like a lot of languages do, that explicitly converts the bytes of a string to their corresponding hex values(along the lines of "StringToHex$()" or "str.ToHex()" or something), then yes, getting the values that you were getting would be incorrect, and that would be a bug in LB. But that's not what dechex$() is. It's in the name: DecHex. DECimal to HEXadecimal. Convert one number from base 10 representation to base 16 representation.

[Carl Gundel]

Nov 5, 2020 12:43:26 GMT -5 Chris Iverson said:

dechex$ works perfectly, and it properly converts any decimal number you give it to the proper corresponding hexadecimal representation.

Agreed.  If dechex$ works according to the documentation, then it's not a bug.  Do you need Base64 features?  I'll consider it, but you can write your own.