hexatridecimal or base 36

[sjordery]

Hi All,

Part of a requirement I have is to store a number in hexatridecimal format (also known as base 36). Do you know if there are any built-in functions/routines in DataStage that will allow me to do this... or will it need to be something bespoke?

Many thanks in advance.

S

[chulett]

I would wager you'll need to roll your own conversion.

Before reading this, I thought it was a typo as 3x16 is 48 not 36... so, sexatrigesimal?

[sud]

I would wager you'll need to roll your own conversion.

Before reading this, I thought it was a typo as 3x16 is 48 not 36... so, sexatrigesimal?

why did you multiply 3 and 16 ? Decimal is base 10, binary is base 2, hexatridecimal or as you correctly pointed sexatrigesimal is base 36(it uses 0-9 and A-Z). Funny names though

[sjordery]

I would wager you'll need to roll your own conversion.

Before reading this, I thought it was a typo as 3x16 is 48 not 36... so, sexatrigesimal?

Spot on!

The most common latinate name for base 36 seems to be hexatridecimal, although sexatrigesimal would arguably be more correct."

From:

http://en.wikipedia.org/wiki/Base_36

Cheers,
S

[sjordery]

Code: Select all

chars = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ"
lenChars = LEN(chars)

;! Check if ok to convert to this base...

If nbase < 2 or nbase > lenChars Then
  Return('')
End 

r = ''
newNumber = ''

;! r contains offset of char that was converted to new base

Loop

  r = MOD(num,nbase)
  newNumber = chars[r+1,1]:newNumber
  num = INT(num/nbase)
  
While num >= nbase Do repeat

;! Final number to convert

Ans = (chars[num+1,1]:newNumber)

This seems to work for me (BASIC routine as I am non-C++!)

Honest critiques welcome

Regards,
S

[ray.wurlod]

There will be some upper limit to the size of numbers you can convert using ordinary arithmetic, depending on your EXACTNUMERIC setting.

For numbers larger than this you can use "string math" (SAdd(), SMui(), SSub() and SDiv() functions). You need to build a "remainder" function out of SDiv() and SSub() functions.

Did you ask WHY there is this requirement?

[chulett]

why did you multiply 3 and 16 ?

Simply because 'hexadecimal' is base 16 and then I did the 'tri' or 'triple' or 'times 3' part.

[shawn_ramsey]

I wrote code way back in Foxpro and Oracle to convert both ways (integer to/from binary through hexatridecimal). If I recall it was a pretty simple program.

Hmm... I wonder if I still have it somewhere.

[sjordery]

Did you ask WHY there is this requirement?

The DS process will be replacing an existing process, and these codes are historically stored as base-36. Presume it is just space saving..

Cheers
S

[sjordery]

There will be some upper limit to the size of numbers you can convert using ordinary arithmetic, ...Premium Stuff

Oh - and thanks very much for this Ray.

Regards,
S

[sud]

why did you multiply 3 and 16 ?

Simply because 'hexadecimal' is base 16 and then I did the 'tri' or 'triple' or 'times 3' part.

Yep but any base conversion is merely positional. There are 0-9 i.e., 10 characters and A-Z 26 more adding to 36 and thats the base 36's significance. The name is 'hexatridecimal' meaning 6(hex) + 3(tri) * 10(decimal) = 36.

Hexadecimal = 6(hex) + 10(decimal) = 16

[chulett]

Wow... you guys are way too anal. Did you not notice Mr Winkie? You know, as in 'wink, wink, nudge, nudge'.

[sud]

Wow... you guys are way too anal. Did you not notice Mr Winkie? You know, as in 'wink, wink, nudge, nudge'.

Lest anyone misunderstands, Craig, one of our most valuable posters, means anal as in analytical.

[chulett]

Umm... yah! That's exactly what I meant.