Convert from numbers to letters

[rh0dium]

Hi All,

While I know there is a zillion ways to do this.. What is the most
efficient ( in terms of lines of code ) do simply do this.

a=1, b=2, c=3 ... z=26

Now if we really want some bonus points..

a=1, b=2, c=3 ... z=26 aa=27 ab=28 etc..

Thanks

[Dan Sommers]

On 19 May 2005 06:56:45 -0700,
"rh0dium" <sk****@pointcircle.com> wrote:

Hi All,
While I know there is a zillion ways to do this.. What is the most
efficient ( in terms of lines of code ) do simply do this. a=1, b=2, c=3 ... z=26
(a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p,q,r,s,t,u,v,w,x,y ,z) = range( 1, 27 )
Now if we really want some bonus points.. a=1, b=2, c=3 ... z=26 aa=27 ab=28 etc..

It's still one line, following the pattern from above, just longer.

Now why do you want to do this?

Regards,
Dan

--
Dan Sommers
<http://www.tombstonezero.net/dan/>

[Bill Mill]

On 19 May 2005 06:56:45 -0700, rh0dium <sk****@pointcircle.com> wrote:

Hi All,

While I know there is a zillion ways to do this.. What is the most
efficient ( in terms of lines of code ) do simply do this.

a=1, b=2, c=3 ... z=26

Now if we really want some bonus points..

a=1, b=2, c=3 ... z=26 aa=27 ab=28 etc..

just for fun, here is one way to do it with a listcomp. Obfuscated
python fans, rejoice!

alpha = 'abcdefghijklmnopqrstuvwxyz'
for i, digraph in enumerate(sorted([''.join((x, y)) for x in alpha \ for y in [''] + [z for z in alpha]], key=len)):
.... locals()[digraph] = i + i
.... a 1 b 2 ac 29 dg 111 zz 702 26**2 + 26

702
Thanks

--
http://mail.python.org/mailman/listinfo/python-list

[Jason Drew]

It seems strange to want to set the values in actual variables: a, b,
c, ..., aa, ab, ..., aaa, ..., ...

Where do you draw the line?

A function seems more reasonable. "In terms of lines of code" here is
my terse way of doing it:

nrFromDg = lambda dg: sum(((ord(dg[x])-ord('a')+1) * (26 **
(len(dg)-x-1)) for x in xrange(0, len(dg))))

Then, for example
nrFromDg("bc")
gives
55
and
nrFromDg("aaa")
gives
703
and so on for whatever you want to evaluate.

This is efficient in terms of lines of code, but of course the function
is evaluating ord("a") and len(dg) multiple times, so it's not the most
efficient in terms of avoiding redundant calculations. And
nrFromDg("A") gives you -31, so you should really force dg into
lowercase before evaluating it. Oh, and it's pretty hard to read that
lambda expression.

"Least amount of code" == "best solution"
False

[Steven Bethard]

Bill Mill wrote:

py> alpha = 'abcdefghijklmnopqrstuvwxyz'
py> for i, digraph in enumerate(sorted([''.join((x, y)) for x in alpha
... for y in [''] + [z for z in alpha]], key=len)):
... locals()[digraph] = i + i
...

It would probably be better to get in the habit of writing
globals()[x] = y
instead of
locals()[x] = y
You almost never want to do the latter[1]. The only reason it works in
this case is because, at the module level, locals() is globals().

You probably already knew this, but I note it here to help any newbies
avoid future confusion.

Steve

[1] For 99% of use cases. Modifying locals() might be useful if you're
just going to pass it to another function as a dict. But I think I've
seen *maybe* 1 use case for this.

[qwweeeit]

Hi rh0dium,
Your request gives me the opportunity of showing a more realistic
example of the technique of "self-modification coding".
Although the coding is not as short as that suggested by the guys who
replayed to you, I think that it can be interesting....

# newVars.py
lCod=[]
for n in range(1,27):
.. lCod.append(chr(n+96)+'='+str(n)+'\n')
# other for-loops if you want define additional variables in sequence
(ex. aa,bb,cc etc...)
# write the variable definitions in the file "varDef.py"
fNewV=open('varDef.py','w')
fNewV.writelines(lCod)
fNewV.close()
from varDef import *
# ...
If you open the generated file (varDef.py) you can see all the variable
definitions, which are runned by "from varDef import *"
Bye.

[rh0dium]

Call me crazy.. But it doesn't work..

for i, digraph in enumerate(sorted([''.join((x, y)) for x in alpha for
y in [''] + [z for z in alpha]], key=len)):
globals()[digraph]=i+1

How do you implement this sucker??

Thanks

[rh0dium]

This is great but backwards...

Ok because you all want to know why.. I need to convert Excel columns
A2 into , [1,0] and I need a simple way to do that..

( The way this works is A->0 and 2->1 -- Yes they interchange -- So
B14 == [13,1] )

So my logic was simple convert the A to a number and then do the swap.
I didn't really care about the function so to speak it was a minor step
in the bigger picture..

By the way if you haven't played with pyXLWriter is it really good :)

So can anyone simply provide a nice function to do this? My logic was
along the same lines as Dans was earlier - but that just seems too
messy (and ugly)

Thanks

[Bill Mill]

On 19 May 2005 11:52:30 -0700, rh0dium <sk****@pointcircle.com> wrote:

Call me crazy.. But it doesn't work..

What doesn't work? What did python output when you tried to do it? It
is python 2.4 specific, it requires some changes for 2.3, and more for
earlier versions of python.
for i, digraph in enumerate(sorted([''.join((x, y)) for x in alpha for
y in [''] + [z for z in alpha]], key=len)):
globals()[digraph]=i+1

How do you implement this sucker??

Works just fine for me. Let me know what error you're getting and I'll
help you figure it out.

Peace
Bill Mill
bill.mill at gmail.com

[Bill Mill]

On 19 May 2005 11:59:00 -0700, rh0dium <sk****@pointcircle.com> wrote:

This is great but backwards...

Ok because you all want to know why.. I need to convert Excel columns
A2 into , [1,0] and I need a simple way to do that..

( The way this works is A->0 and 2->1 -- Yes they interchange -- So
B14 == [13,1] )

why didn't you say this in the first place?

def coord2tuple(coord):
row, col = '', ''
alpha = 'abcdefghijklmnopqrstuvwxyz'.upper()
pairs = [''.join((x,y)) for x in alpha for y in [''] + [z for z in alpha]]
pairs = sorted(pairs, key=len)
coord = coord.upper()
for c in coord:
if c in alpha:
row += c
else:
col += c
return (int(col)-1, pairs.index(row))

coord2tuple('B14') (13, 1) coord2tuple('ZZ14') (13, 701) coord2tuple('ZZ175') (174, 701) coord2tuple('A2')

(1, 0)

Are there cols greater than ZZ? I seem to remember that there are not,
but I could be wrong.

Hope this helps.

Peace
Bill Mill
bi*******@gmail.com

[rh0dium]

Python 2.3.5 (#1, Mar 20 2005, 20:38:20)
[GCC 3.3 20030304 (Apple Computer, Inc. build 1809)] on darwin

Traceback (most recent call last):
File "<stdin>", line 1, in ?
NameError: name 'sorted' is not defined

I think you're probably using 2.4 ??

[Bill Mill]

On 19 May 2005 12:20:03 -0700, rh0dium <sk****@pointcircle.com> wrote:

Python 2.3.5 (#1, Mar 20 2005, 20:38:20)
[GCC 3.3 20030304 (Apple Computer, Inc. build 1809)] on darwin

Traceback (most recent call last):
File "<stdin>", line 1, in ?
NameError: name 'sorted' is not defined

I think you're probably using 2.4 ??
Yes, sorted() is new in python 2.4 .You could use a very lightly
tested pure-python partial replacement:

def sorted(lst, **kwargs):
l2 = lst[:]
if kwargs.has_key('key'):
f = kwargs['key']
l2.sort(lambda a,b: cmp(f(a), f(b)))
return l2
l2.sort()
return l2

And from your other email: I need to go the other way! tuple2coord

Sorry, I only go one way. It should be transparent how to do it backwards.

Peace
Bill Mill
bi*******@gmail.com

[Peter Otten]

Bill Mill wrote:

Traceback (most recent call last):
File*"<stdin>",*line*1,*in*?
NameError: name 'sorted' is not defined

I think you're probably using 2.4 ??

Yes, sorted() is new in python 2.4 .You could use a very lightly
tested pure-python partial replacement:

By the way, sorted() can be removed from your original post.

Code has no effect :-)

Peter

[Bill Mill]

On 5/19/05, Peter Otten <__*******@web.de> wrote:

Bill Mill wrote:

Traceback (most recent call last):
File"<stdin>",line1,in?
NameError: name 'sorted' is not defined

I think you're probably using 2.4 ??

Yes, sorted() is new in python 2.4 .You could use a very lightly
tested pure-python partial replacement:

By the way, sorted() can be removed from your original post.

Code has no effect :-)

I'm gonna go ahead and disagree with you:

sorted([''.join((x, y)) for x in alpha \ .... for y in [''] + [z for z in alpha]], key=len) == \
.... [''.join((x,y)) for x in alpha for y in [''] + [z for z in alpha]]
False

If you want to see why, here's a small example:
alpha = 'abc'
[''.join((x,y)) for x in alpha for y in [''] + [z for z in alpha]] ['a', 'aa', 'ab', 'ac', 'b', 'ba', 'bb', 'bc', 'c', 'ca', 'cb', 'cc']
sorted([''.join((x,y)) for x in alpha for y in [''] + [z for z in alpha]],

key=len)
['a', 'b', 'c', 'aa', 'ab', 'ac', 'ba', 'bb', 'bc', 'ca', 'cb', 'cc']

Peace
Bill Mill
bill.mill at gmail.com

[Gary Wilson Jr]

Bill Mill wrote:

On 5/19/05, Peter Otten <__*******@web.de> wrote:

Bill Mill wrote:

Traceback (most recent call last):
File"<stdin>",line1,in?
NameError: name 'sorted' is not defined

I think you're probably using 2.4 ??

Yes, sorted() is new in python 2.4 .You could use a very lightly
tested pure-python partial replacement:
By the way, sorted() can be removed from your original post.

Code has no effect :-)

I'm gonna go ahead and disagree with you:

Me too, although I would forgo the sort altogether (while making things
a little more readable IMO):
alpha = 'abcdefghijklmnopqrstuvwxyz'.upper()
pairs = [''.join((x,y)) for x in alpha for y in [''] + [z for z in alpha]]
pairs = sorted(pairs, key=len)

alpha = 'abcdefghijklmnopqrstuvwxyz'.upper()
pairs = [x for x in alpha] + [''.join((x,y)) for x in alpha for y in alpha]

[Peter Otten]

Bill Mill wrote:

By the way, sorted() can be removed from your original post.

Code has no effect :-)

I'm gonna go ahead and disagree with you:

sorted([''.join((x, y)) for x in alpha \ ...****for*y*in*['']*+*[z*for*z*in*alpha]],*key=len)*==*\
... [''.join((x,y)) for x in alpha for y in [''] + [z for z in alpha]]
False

That's not your original code. You used the contents to modify the locals()
(effectively globals()) dictionary:

alpha = 'abcdefghijklmnopqrstuvwxyz'
for i, digraph in enumerate(sorted([''.join((x, y)) for x in alpha \ for*y*in*['']*+*[z*for*z*in*alpha]],*key=len)):
...*****locals()[digraph]*=*i*+*i
...

Of course you lose the order in that process.
When you do care about order, I suggest that you swap the for clauses
instead of sorting, e. g:

alpha = list("abc")
items = [x + y for x in [""] + alpha for y in alpha]
items == sorted(items, key=len)

True

Peter

[Peter Otten]

Peter Otten wrote:

[Something stupid]

You are right. I finally got it.

Peter

[Jason Drew]

We weren't really backwards; just gave a full solution to a half-stated
problem.

Bill, you've forgotten the least-lines-of-code requirement :-)

Mine's still a one-liner (chopped up so line breaks don't break it):

z = lambda cp: (int(cp[min([i for \
i in xrange(0, len(cp)) if \
cp[i].isdigit()]):])-1,
sum(((ord(cp[0:min([i for i in \
xrange(0, len(cp)) if \
cp[i].isdigit()])][x])-ord('A')+1) \
* (26 ** (len(cp[0:min([i for i in \
xrange(0, len(cp)) if \
cp[i].isdigit()])])-x-1)) for x in \
xrange(0, len(cp[0:min([i for i in \
xrange(0, len(cp)) if \
cp[i].isdigit()])]))))-1)

print z("B14")
# gives (13, 1)

Maybe brevity isn't the soul of wit after all ...

[Mike Meyer]

Bill Mill <bi*******@gmail.com> writes:

On 19 May 2005 11:59:00 -0700, rh0dium <sk****@pointcircle.com> wrote:

This is great but backwards...

Ok because you all want to know why.. I need to convert Excel columns
A2 into , [1,0] and I need a simple way to do that..

( The way this works is A->0 and 2->1 -- Yes they interchange -- So
B14 == [13,1] )

why didn't you say this in the first place?

def coord2tuple(coord):
row, col = '', ''
alpha = 'abcdefghijklmnopqrstuvwxyz'.upper()
pairs = [''.join((x,y)) for x in alpha for y in [''] + [z for z in alpha]]
pairs = sorted(pairs, key=len)
coord = coord.upper()
for c in coord:
if c in alpha:
row += c
else:
col += c
return (int(col)-1, pairs.index(row))

That seems like the long way around. Python can search strings for
substrings, so why not use that? That gets the search loop into C
code, where it should be faster.

from string import uppercase

def coord2tuple2(coord):
if len(coord) > 1 or uppercase.find(coord) < 0:
raise ValueError('coord2tuple2 expected a single uppercase character, got "%s"' % coord)
return uppercase.index(coord) + 1

Without the initial test, it has a buglet of return values for "AB"
and similar strings. If searching uppercase twice really bothers you,
you can drop the uppercase.find; then you'll get less informative
error messages if coord2tuple2 is passed single characters that aren't
in uppercase.

<mike
--
Mike Meyer <mw*@mired.org> http://www.mired.org/home/mwm/
Independent WWW/Perforce/FreeBSD/Unix consultant, email for more information.

[Gary Wilson Jr]

Gary Wilson Jr wrote:

alpha = 'abcdefghijklmnopqrstuvwxyz'.upper()
pairs = [x for x in alpha] + [''.join((x,y)) for x in alpha for y in alpha]

I forget, is string concatenation with '+' just as fast as join()
now (because that would look even nicer)?

[Steven Bethard]

Jason Drew wrote:

z = lambda cp: (int(cp[min([i for \
i in xrange(0, len(cp)) if \
cp[i].isdigit()]):])-1,
sum(((ord(cp[0:min([i for i in \
xrange(0, len(cp)) if \
cp[i].isdigit()])][x])-ord('A')+1) \
* (26 ** (len(cp[0:min([i for i in \
xrange(0, len(cp)) if \
cp[i].isdigit()])])-x-1)) for x in \
xrange(0, len(cp[0:min([i for i in \
xrange(0, len(cp)) if \
cp[i].isdigit()])]))))-1)

While I think we can all agree that this is a sin against man and nature
;) I'll ignore that for the moment to note that you don't need any of
the '\' characters. You're already using parentheses and brackets. I
find there are *very* few cases where I really need a line-continuation
character.

STeVe

[Steven Bethard]

Gary Wilson Jr wrote:

Gary Wilson Jr wrote:

alpha = 'abcdefghijklmnopqrstuvwxyz'.upper()
pairs = [x for x in alpha] + [''.join((x,y)) for x in alpha for y in alpha]

I forget, is string concatenation with '+' just as fast as join()
now (because that would look even nicer)?

Certain looping constructs like:

x = ''
for y in z:
x += y

are now (in CPython 2.4) somewhere near the speed of:

x = []
for y in z:
x.append(y)
x = ''.join(x)

but this isn't really relevant to your problem because you're only
joining two characters:

$ python -m timeit -s "import string; a = string.ascii_uppercase"
"[''.join([x, y]) for x in a for y in a]"
1000 loops, best of 3: 1.02 msec per loop

$ python -m timeit -s "import string; a = string.ascii_uppercase"
"[x + y for x in a for y in a]"
1000 loops, best of 3: 295 usec per loop

Unsurprisingly, it's actually faster to simply concatenate the two
characters.

STeVe

[Jason Drew]

Oh yeah, oops, thanks. (I mean the line continuations, not the alleged
sin against man and nature, an accusation which I can only assume is
motivated by jealousy :-) Or fear? They threw sticks at Frankenstein's
monster too. And he turned out alright.

My elegant "line" of code started out without the enclosing
parentheses; forgot I didn't need the \s when I embraced it.

[rh0dium]

Wow - now that is ugly.. But it is effective. I would love a cleaner
version - but I did say brevity :)

Nice work.

[rh0dium]

Now can you reverse this process tuple2coord??

Thats what I'm really after :)

[Mike Meyer]

"rh0dium" <sk****@pointcircle.com> writes:

Now can you reverse this process tuple2coord??

You didn't provide enough context to know who you're asking, but
here's the inverse of my coord2tuple2 function:

from string import uppercase
def tuple2coord(number):
if 1 > number or number > 26:
raise ValueError("tuple2coord expected a number between 1 and 26, got '%s'" % number)
return (" " + uppercase)[number]

<mike
--
Mike Meyer <mw*@mired.org> http://www.mired.org/home/mwm/
Independent WWW/Perforce/FreeBSD/Unix consultant, email for more information.

[Jason Drew]

Er, yes! It's REALLY ugly! I was joking (though it works)! I retract it
from the code universe. (But patent pending nr. 4040404.)

Here's how I really would convert your (row_from_zero, col_from_zero)
tuple to spreadsheet "A1" coords, in very simple and easy to read code.

##def tuple2coord(tupl):
## def colnr2digraph(colnr):
## if colnr <= 26:
## return chr(ord('A') + colnr-1)
## m = colnr % 26
## if m == 0:
## m = 26
## h = (colnr - m) / 26
## return colnr2digraph(h) + colnr2digraph(m)
##
## rowfromzero, colfromzero = tupl
## row = rowfromzero+1
## col = colfromzero+1
## return colnr2digraph(col) + str(row)
##
##print tuple2coord((13,702))
### gives AAA14
### (because the tuple counts rows and columns from zero)

Note that this allows column nrs of any size, not just up to "ZZ". If
you really know the column limit is ZZ, then a lookup dictionary would
be a more efficient speed-wise solution. (Though I'd still use my nice
recursive no-brainer colnr2digraph function to populate the
dictionary.)

P.S. the line that says
h = (colnr - m) / 26
could really, in current Python, be just
h = colnr / 26
but the former is more language- and future-neutral.

[Jason Drew]

Sorry, scratch that "P.S."! The act of hitting Send seems to be a great
way of realising one's mistakes.

Of course you need colnr - m for those times when m is set to 26.
Remembered that when I wrote it, forgot it 2 paragraphs later!

[Steven Bethard]

Jason Drew wrote:

##def tuple2coord(tupl): [snip] ## rowfromzero, colfromzero = tupl

Just a side note here that if you want a better function signature, you
might consider writing this as:

tuple2coord((rowfromzero, colfromzero)):
...

Note that the docstrings are nicer this way:

py> def tuple2coord(tupl):
.... x, y = tupl
....
py> help(tuple2coord)
Help on function tuple2coord in module __main__:

tuple2coord(tupl)

py> def tuple2coord((x, y)):
.... pass
....
py> help(tuple2coord)
Help on function tuple2coord in module __main__:

tuple2coord((x, y))

STeVe

[Jason Drew]

Hey, that's good. Thanks Steve. Hadn't seen it before. One to use.

Funny that Pythonwin's argument-prompter (or whatever that feature is
called) doesn't seem to like it.

E.g. if I have
def f(tupl):
print tupl

Then at the Pythonwin prompt when I type
f(
I correctly get "(tupl)" in the argument list pop-up box.

But if I have
def f((a, b)):
print a, b

then when I type
f(
I just get "(.0)" in the argument list pop-up box.

Or with
def f(p, q, (a, b)):
pass
Pythonwin prompts with
"(p, q, .4)"
However in each case the help() function correctly lists all the
arguments. Strange. I'll check if it's a known "feature".

This is with
"PythonWin 2.4 (#60, Feb 9 2005, 19:03:27) [MSC v.1310 32 bit (Intel)]
on win32."

[Bill Mill]

On 20 May 2005 10:07:55 -0700, Jason Drew <ja*********@gmail.com> wrote:

Hey, that's good. Thanks Steve. Hadn't seen it before. One to use.

Funny that Pythonwin's argument-prompter (or whatever that feature is
called) doesn't seem to like it.

E.g. if I have
def f(tupl):
print tupl

Then at the Pythonwin prompt when I type
f(
I correctly get "(tupl)" in the argument list pop-up box.

But if I have
def f((a, b)):
print a, b

then when I type
f(
I just get "(.0)" in the argument list pop-up box.

Or with
def f(p, q, (a, b)):
pass
Pythonwin prompts with
"(p, q, .4)"


However in each case the help() function correctly lists all the
arguments. Strange. I'll check if it's a known "feature".
That sounds like a bug in pythonwin autocomplete. Tuple unpacking in
function arguments is definitely a known feature, there were some
recent (fairly extensive) clp threads about it.[1]

I wish people would use it more, I think it's an awesome feature when
properly used. I like it especially for signatures like "def
change_coord((x, y))". It was one of those features, for me, where I
just tried it without knowing of its existence, assuming it would
work, and I was pleasantly surprised that it did.

Peace
Bill Mill
bill.mill at gmail.com

[1] http://tinyurl.com/89zar

I think there was another about ways to improve tuple unpacking, but I
didn't find it in a brief search.

This is with
"PythonWin 2.4 (#60, Feb 9 2005, 19:03:27) [MSC v.1310 32 bit (Intel)]
on win32."

--
http://mail.python.org/mailman/listinfo/python-list