Prev: xargs rm -rf in crontab
Next: Dir Files & Size
From: david_kirby_byers on 5 Jul 2006 12:21
is there a simply way to get the sum of a column im trying to get a
average?
so suggestions would be appreciated.

below is my column from that i would like to then see what percent 4765
is of the total sum of the column.

4765
1876
1107
1088
883
670
420
335
179
177
136
129
126
120
120
110
77
66
60
51
45
22
17
17
12
11
10
8
6
6
4
3
3
3
3
1
0

From: Xicheng Jia on 5 Jul 2006 12:54
david_kirby_byers(a)yahoo.com wrote:
> is there a simply way to get the sum of a column im trying to get a
> average?
> so suggestions would be appreciated.
>
> below is my column from that i would like to then see what percent 4765
> is of the total sum of the column.
>

a FAQ:-)

awk '{x+=$1}END{print 4765/x}' file
perl -pe '$\+=$_}{$\=4765/$\' file
perl -lne '$x+=$_}{print 4765/$x' file

Xicheng

From: david_kirby_byers on 5 Jul 2006 12:56
paste -s file | sed 's/ / + /g' | bc

heres what i have came up with to add column .....


david_kirby_byers(a)yahoo.com wrote:
> is there a simply way to get the sum of a column im trying to get a
> average?
> so suggestions would be appreciated.
>
> below is my column from that i would like to then see what percent 4765
> is of the total sum of the column.
>
> 4765
> 1876
> 1107
> 1088
> 883
> 670
> 420
> 335
> 179
> 177
> 136
> 129
> 126
> 120
> 120
> 110
> 77
> 66
> 60
> 51
> 45
> 22
> 17
> 17
> 12
> 11
> 10
> 8
> 6
> 6
> 4
> 3
> 3
> 3
> 3
> 1
> 0

From: Stephane CHAZELAS on 5 Jul 2006 13:15
2006-07-5, 09:56(-07), david_kirby_byers(a)yahoo.com:
> paste -s file | sed 's/ / + /g' | bc

paste -sd+ file | bc

--
St?phane
From: Penguiniator on 5 Jul 2006 13:35
david_kirby_byers(a)yahoo.com wrote:


> is there a simply way to get the sum of a column im trying to
> get a average?

If your data is in a file named data and the data is in column
one:

sum=0
count=0
for i in $(awk '{print $1}' data);do
sum=$(( $sum + $i ))
count=$(( $count + 1 ))
done
echo $(( $sum / $count ))

Or you can use bc for more precise results:
echo "scale=2; $sum / $count" | bc

where scale is the number of significant digits.

Also, you can pipe the data from standard input with:
.... | for i in $(echo $(cat) | tr ' ' '\n' | awk ...); do
...
done

The previous works when there is only one column of numbers in
your data file. Otherwise, you will have to isolate that column
first. You can use awk or cut for that. The tr in the pipeline
is necessary. Otherwise, cat removes the newlines and awk sees a
single record containing all your data and prints only the first
field; i.e., only the first number in the data stream.

And you can combine these techniques to take input from a file
specified as a parameter or from redirected standard input with
simple tests such as: [ -z "$1" ] && { ...; } || { ...; }
or: if [ -z "$1" ]; then ...; else ...; fi;

where $1 is, or is not, an empty string corresponding to the
filename containing your data.

> so suggestions would be appreciated.
>
> below is my column from that i would like to then see what
> percent 4765 is of the total sum of the column.
>
> 4765
> ....

This seems to be a slightly different problem from the first
question. And the answer would depend somewhat on what 4765
represents. Is it a fixed value? Is it the last value? Is it the
first value? Is it the most recent value? Or is it simply the
biggest value? Determining that will help determine how to
isolate it from the other values in the column.

biggest value?: biggest=$(sort -gr | head -1)
first value?: first=$(head -1)
last value?: last=$(tail -1)
fixed value?: fixed=4765

After that, the calculation is simple:
echo $(( 4765 / $sum * 100 ))

substituting bc as above if desired and assuming that by
"average" you mean "mean".

Hope this helps.
 |  Next  |  Last
Pages: 1 2
Prev: xargs rm -rf in crontab
Next: Dir Files & Size