ways to put data into arrays

I have a list (see below) I think should be stored in arrays.

However, it seems it will be easier to simply use NR to record each line instead of trying to get a multidimensional array simulation so that each month gets all days associated with it, perhaps like :

01_01_2012
02_01_2012
14_01_2012

....etc.

example day/month/year list:
01 01 2012
02 01 2012
14 01 2012
22 01 2012
02 02 2012
03 02 2012
....etc.

Other suggestions appreciated.

-Bryan

On 2/26/2012 11:39 AM, bryanlepore@gmail.com wrote:
> I have a list (see below) I think should be stored in arrays.

Why?

> However, it seems it will be easier to simply use NR to record each line instead of trying to get a multidimensional array simulation so that each month gets all days associated with it, perhaps like :
>
> 01_01_2012
> 02_01_2012
> 14_01_2012
>
> ...etc.
>
> example day/month/year list:
> 01 01 2012
> 02 01 2012
> 14 01 2012
> 22 01 2012
> 02 02 2012
> 03 02 2012
> ...etc.
>
> Other suggestions appreciated.

It's very difficult to suggest a solution as you haven't described the problem,
just told us what you think might be a solution.

If you post some small sample input set with the expected output from that input
and a brief description of what you're trying to do, THEN you can get a sensible
suggestion for a solution. Otherwise you'd just be getting vague guesses that
may be totally inappropriate for your problem.

Ed.

On Feb 26, 1:09*pm, Ed Morton <mortons...@gmail.com> wrote:
> post some small sample input set with the expected output from that input
> and a brief description of what you're trying to do

this is changing every time I think about it, but data file :

01 01 2012 1234.23
02 01 2012 3.23
02 01 2012 45.11
14 01 2012 23.51
22 01 2012 526.67
22 01 2012 5.78
02 02 2012 1.67
03 02 2012 -3.69

I think I want to make one array per month ($2) that contains each day
($1) associated with each value $4. AFAIK at least arrays in awk can/
should do that, though I have considered using a more appropriate
program. but this looks very amenable to awk.

-Bryan

On 2/26/2012 12:30 PM, B wrote:
> On Feb 26, 1:09 pm, Ed Morton<mortons...@gmail.com> wrote:
>> post some small sample input set with the expected output from that input
>> and a brief description of what you're trying to do

sample input - posted
expected output - not posted
brief description - not posted

>
> this is changing every time I think about it, but data file :
>
> 01 01 2012 1234.23
> 02 01 2012 3.23
> 02 01 2012 45.11
> 14 01 2012 23.51
> 22 01 2012 526.67
> 22 01 2012 5.78
> 02 02 2012 1.67
> 03 02 2012 -3.69
>
> I think I want to make one array per month ($2) that contains each day
> ($1) associated with each value $4. AFAIK at least arrays in awk can/
> should do that, though I have considered using a more appropriate
> program. but this looks very amenable to awk.
>
> -Bryan

Again, though, you're not telling us WHAT you're trying to do you're just
telling us HOW you think you should do it.

I can easily say that to the solution you describe above could be implemented as:

{ arr[$2] = arr[$2] $1 " " $4 ";" }

but I have absolutely no idea if that solution is a good way to solve your
problem or not as I still don't know what the problem actually IS that you're
trying to solve.

If you can't at least post some expected output from the above input and a brief
description of why that's the output, then it's not worth your while posting
here as the answers you get could be totally wrong for your actual application.

Ed.

On 2/26/2012 12:52 PM, Ed Morton wrote:
> On 2/26/2012 12:30 PM, B wrote:
>> On Feb 26, 1:09 pm, Ed Morton<mortons...@gmail.com> wrote:
>>> post some small sample input set with the expected output from that input
>>> and a brief description of what you're trying to do
>
> sample input - posted
> expected output - not posted
> brief description - not posted
>
>>
>> this is changing every time I think about it, but data file :
>>
>> 01 01 2012 1234.23
>> 02 01 2012 3.23
>> 02 01 2012 45.11
>> 14 01 2012 23.51
>> 22 01 2012 526.67
>> 22 01 2012 5.78
>> 02 02 2012 1.67
>> 03 02 2012 -3.69
>>
>> I think I want to make one array per month ($2) that contains each day
>> ($1) associated with each value $4. AFAIK at least arrays in awk can/
>> should do that, though I have considered using a more appropriate
>> program. but this looks very amenable to awk.
>>
>> -Bryan
>
> Again, though, you're not telling us WHAT you're trying to do you're just
> telling us HOW you think you should do it.
>
> I can easily say that to the solution you describe above could be implemented as:
>
> { arr[$2] = arr[$2] $1 " " $4 ";" }
>
> but I have absolutely no idea if that solution is a good way to solve your
> problem or not as I still don't know what the problem actually IS that you're
> trying to solve.
>
> If you can't at least post some expected output from the above input and a brief
> description of why that's the output, then it's not worth your while posting
> here as the answers you get could be totally wrong for your actual application.
>
> Ed.

Maybe this will help. The following may or may not be applicable to your actual
problem, it's just an example of something you might be trying to do to show you
how to describe what you actually ARE trying to do:

Description:
I need to take a list of dates + amounts and produce a list per date (one per
line) of the amounts associated with that date with the amounts separated by
tabs. The output does not need to be sorted by date.

Sample Input:
01 01 2012 1234.23
02 01 2012 3.23
02 01 2012 45.11
14 01 2012 23.51
22 01 2012 526.67
22 01 2012 5.78
02 02 2012 1.67
03 02 2012 -3.69

Expected Output:
2012 01 01 1234.23
2012 01 02 3.23 45.11
2012 01 14 23.51
2012 01 22 526.67 5.78
2012 02 02 1.67
2012 02 03 -3.69


Given the above I'd reply with something like this:

$ cat tst.awk
{
date = $3 " " $2 " " $1
amount = $4
dt2amt[date] = dt2amt[date] "\t" amount
}
END {
for (date in dt2amt) {
print date dt2amt[date]
}
}

$ awk -f tst.awk file
2012 02 03 -3.69
2012 01 14 23.51
2012 01 01 1234.23
2012 01 02 3.23 45.11
2012 02 02 1.67
2012 01 22 526.67 5.78

but again it all depends what you're actually trying do DO, not how you think
you should do it.

Ed.

Ed, the examples you wrote appear to be what is called concatenation
-- I was entirely unaware of how this works though I had heard about
it. The output appears very interesting, because it has each cost
associated with a year, day and date. Then I ought to be able to do a
lot of math and sorting on it in various ways.

I will have to work on these excellent suggestions and the other
interesting recent threads.

Thanks a lot for the help, I appreciate it, especially since I know my
question is rather nebulous.

-Bryan

p.s: time for me to read Eric Raymond's "How To Ask Smart Questions"
again ...

On Sunday, February 26, 2012 2:18:07 PM UTC-5, Ed Morton wrote:
> $ cat tst.awk
> {
> date = $3 " " $2 " " $1
> amount = $4
> dt2amt[date] = dt2amt[date] "\t" amount
> }
> END {
> for (date in dt2amt) {
> print date dt2amt[date]
> }
> }
>
> $ awk -f tst.awk file
> 2012 02 03 -3.69
> 2012 01 14 23.51
> 2012 01 01 1234.23
> 2012 01 02 3.23 45.11
> 2012 02 02 1.67
> 2012 01 22 526.67 5.78
>
> but again it all depends what you're actually trying do DO

IIUC

dt2amt[date] in this sample contains concatenated strings, not variables, and therefore arithmetic operators will not work on them. this is a consequence of awk not providing a concatenation operator as such.

.... I would really like to add those numbers together e.g. on a per-day basis.

-Bryan

bryanlepore@gmail.com <bryanlepore@gmail.com> wrote:

> On Sunday, February 26, 2012 2:18:07 PM UTC-5, Ed Morton wrote:
> > $ cat tst.awk
> > {
> > date = $3 " " $2 " " $1
> > amount = $4
> > dt2amt[date] = dt2amt[date] "\t" amount
> > }
> > END {
> > for (date in dt2amt) {
> > print date dt2amt[date]
> > }
> > }
> >
> > $ awk -f tst.awk file
> > 2012 02 03 -3.69
> > 2012 01 14 23.51
> > 2012 01 01 1234.23
> > 2012 01 02 3.23 45.11
> > 2012 02 02 1.67
> > 2012 01 22 526.67 5.78
> >
> > but again it all depends what you're actually trying do DO
>
> IIUC
>
> dt2amt[date] in this sample contains concatenated strings, not variables, and
therefore arithmetic operators will not work on them. this is a consequence of
awk not providing a concatenation operator as such.

Sorry I don't understand you. Arrays hold data. What would it mean for an array
to hold a variable? What difference would it make to the data if you needed and
explicit operator to concatenate strings?

>
> .... I would really like to add those numbers together e.g. on a per-day basis.

Then just use add the values as numbers instead of concatenating them as strings:

{
date = $3 " " $2 " " $1
amount = $4
dt2amt[date] += amount
}
END {
for (date in dt2amt) {
print date, dt2amt[date]
}
}

Regards,

Ed.

Posted using www.webuse.net