## Average columns based on header name

# 1
Average columns based on header name

Hi Friends,

I have files with columns like this. This sample input below is partial.

Please check below for main file link. Each file will have only two rows.

Basically, here is what I need to be done.

a. Start from second column which is 0.4% here.
b. Go until you hit "10" in the header name. If the header name is exactly 10.0%, then include that column too. If not, only include until the column before it. In this example, since we have 10.1% (29th column), we will be including columns starting from 0.4%(second) until 9.8% which is the 28th column. If the 29th column was to be 10.0%, then it would have been included too.
c. Average the values for these respective columns in the second row (data is not presented here - please click this link for total dataset - https://goo.gl/W8jND7). In this example, starting from 0.4%(second column) till 9.8%(28th column).
d. In the output, print first column which is "Gene", and this average value with column header being

e. Then start from 10.1% (29th column) and check until you hit "20" in the header name. Repeat steps b through d. And print output as

Repeat this until you have

f. After you hit 100%, it means one dataset is done.

g. If you observe my column header carefully here, there is another 0.4%-100% columns after the first 100%. I will be having 13 of these 0.4%-100%s in the input file at the above link.

i. I have multiple files, the headers can be

It varies from file to file. But the logic of averaging(if you hit "10", "20", etc) is always the same. And the number of samples 13 is also same which means each file will have 100%s for 13 times.

P.S: A Bonus of 1000 bits will be awarded to the effectively working solution.
 jacobs.smith View Public Profile for jacobs.smith Find all posts by jacobs.smith
# 2
I do not understand what output you are trying to produce.

For each input gene line do you want:
1. one output line with the gene number and 10 averages from all 13 samples,
2. thirteen output lines with the gene number and the 10 averages from one sample on each line, or
3. one output line with the gene number and 130 averages where each set of 10 averages comes from one sample?
Can you show us the exact output you're hoping to produce from the data provided in your sample input for genes 1 and 2?

Was the data for gene 3 in your sample truncated, or will some inputs have missing fields that should be treated as zero values when computing the averages?
This User Gave Thanks to Don Cragun For This Post:
 Don Cragun View Public Profile for Don Cragun Find all posts by Don Cragun
# 3
Quote:
Originally Posted by Don Cragun
I do not understand what output you are trying to produce.

For each input gene line do you want:
1. one output line with the gene number and 10 averages from all 13 samples,
2. thirteen output lines with the gene number and the 10 averages from one sample on each line, or
3. one output line with the gene number and 130 averages where each set of 10 averages comes from one sample?
Can you show us the exact output you're hoping to produce from the data provided in your sample input for genes 1 and 2?

Was the data for gene 3 in your sample truncated, or will some inputs have missing fields that should be treated as zero values when computing the averages?
Hi Don,

Thank you for your response. Good questions and I am glad at least you replied. Thank you so much.

Here are the answers for your questions.

I want "one output line with the gene number and 130 averages where each set of 10 averages comes from one sample"

Since the data is very big, I will use the below small example assuming I have only one sample.

Here the 2.6 is coming from averaging 1 in second column through 1 in 11th column. 5.9 is coming from averaging from 22.5(12th column) till 7 (21st column). Please remember that if the 11th column was to be any value greater than 10.0% like 10.1% or 10.8% or anything, then we will be averaging only until the 10th column.

The main input file from the link that I gave in my earlier post, has 13 samples' 1% to 100% values and only two rows. In the output file, we will have Gene column plus 130 average values. Makes sense?

There will be no gene3 in an input file. It will always be 2 rows (first line is header and second line is the values for averaging) with 13 sample being chopped into different varying percentages. There will be no missing fields. For every column header, there is a value associated with it in the input file.

Also I have a batch of files like 4000. I give a folder with the extension, and the script should be reading each file and following the averaging conditions of differentiating exact 10.0% (to consider for average) and any other values greater than 10.0% like 10.1% or 10.2% or 10.3% etc (to not consider this column and the column before it).

Please ask as many questions as possible and I will be glad to answer.

Coming to what I have tried so far, I have been trying to read the headers and print each set into a different file and then do the computation and put it back and then move to the other sample. This seems to be very time taking.

All your time and understanding is highly appreciated.

 jacobs.smith View Public Profile for jacobs.smith Find all posts by jacobs.smith
# 4
This is a hell of a problem, by sheer data volume. I'm brain dead now; can't verify the results are correct although I think/hope they are. Try Please do some validation and come back with its result.
This User Gave Thanks to RudiC For This Post:
 RudiC View Public Profile for RudiC Find all posts by RudiC
# 5
Quote:
Originally Posted by RudiC
This is a hell of a problem, by sheer data volume. I'm brain dead now; can't verify the results are correct although I think/hope they are. Try Please do some validation and come back with its result.
Rudic,

You are one piece of a beast! The solution is rocking!

Two quick requests.

1. You are printing the sum and not averages. I just checked and your script is working fine. I am scared to touch this script. Can I print SUM/NR to get the average?

2. Can you make this oneliner? Please...please...please. Again, I can make it. But I don't want to F*** this up. A BIG THANKS MY FRIEND.

When Don asked me how to print the output file, I asked him to print me 130 columns, but your solution seems to be perfectly suitable to my further downstream analysis.

Thank you so much.
 jacobs.smith View Public Profile for jacobs.smith Find all posts by jacobs.smith
# 6
As I said, I'm braindead. I missed the average, sorry. Try printing SUM/(L[i+1]-L[i]). That's not OK for the very first entry as the first field is summed, too, as its value is zero, so the element count is too high. But - I'm scared to touch it, either.

No one liner. Whereever a oneliner can be accommodated, a multiliner will fit, too. Use a function. Source a file.

---------- Post updated at 22:40 ---------- Previous update was at 22:28 ----------

Try this one - the field 1 error should be removed. Please perform a careful validation of the results and come back: ---------- Post updated at 23:39 ---------- Previous update was at 22:40 ----------

Actually, transposing the file like simplified the logics seriously. Piping the paste result into gives you an idea - results may still need a bit polishing.
This User Gave Thanks to RudiC For This Post:
 RudiC View Public Profile for RudiC Find all posts by RudiC
# 7
Hi Rudic,

For some of the files, I am seeing this error with the first long code

Any thoughts?

But the paste tweak is perfect!

Last edited by jacobs.smith; 03-17-2016 at 10:59 PM..
 jacobs.smith View Public Profile for jacobs.smith Find all posts by jacobs.smith

|

## Average of a columns from three files

hello, I have three files in the following order ==> File1 <== 1 20977000 20977000 A C 1.00 0,15 15 45 1 115829313 115829313 G A 0.500 6,7 13 99 ==> File2 <== 1 20977000 20977000 A C 1.00 0,13 13 39 1 115829313 ...

## Find columns in a file based on header and print to new file

Hello, I have to fish out some specific columns from a file based on the header value. I have the list of columns I need in a different file. I thought I could read in the list of headers I need, # file with header names of required columns in required order headers_file=\$2 # read contents...

## Average across multiple columns - awk

Hi forum members, I'm trying to get an average of multiple columns in a csv file using awk. A small example of my input data is as follows: cu,u3o8,au,ag -9,20,-9,3.6 0.005,30,-9,-9 0.005,50,10,3.44 0.021,-9,8,3.35 The following code seems to do most of what I want gawk -F","...

## Keep only columns in first two rows based on partial header pattern.

I have this code below that only prints out certain columns from the first two rows (doesn’t affect rows 3 and beyond). How can I do the same on a partial header pattern “G_TP” instead of having to know specific column numbers (e.g. 374-479)? I’ve tried many other commands within this pipe with no...

## Average across multiple columns group by

Hi experts, I want to group by average, for multiple columns starting column \$7 until NF, group by (\$1-\$5), please help For just 7th column, I can do awk ' NR>1{ arr += \$7 count += 1 } END{ for (a in arr) { print a, arr/count ...

## Make copy of text file with columns removed (based on header)

Hello, I have some tab delimited text files with a three header rows. The headers look like, (sorry the tabs look so messy). index group Name input input input input input input input input input input input...

## Extract columns based on header

Hi to all, I have two files. File1 has no header, two columns: sample1 A sample2 B sample3 B sample4 C sample5 A sample6 D sample7 D File2 has a header, except for the first 3 columns (chr,start,end). "sample1" is the header for the 4th ,5th ,6th columns, "sample2" is the header...

## Average, min and max in file with header, using awk

Hi, I have a file which looks like this: FID IID MISS_PHENO N_MISS N_GENO F_MISS 12AB43131 12AB43131 N 17774 906341 0.01961 65HJ87451 65HJ87451 N 10149 906341 0.0112 43JJ21345 43JJ21345 N 2826 906341 0.003118I would...

## Average of columns with values of other column with same name

I have a lot of input files that have the following form: Sample Cq Sample Cq Sample Cq Sample Cq Sample Cq 1WBIN 23.45 1WBIN 23.45 1CVSIN 23.96 1CVSIN 23.14 S1 31.37 1WBIN 23.53 1WBIN 23.53 1CVSIN 23.81 1CVSIN 23.24 S1 31.49 1WBIN 24.55 1WBIN 24.55 1CVSIN 23.86 1CVSIN 23.24 S1 31.74 ...

## awk based script to find the average of all the columns in a data file

Hi All, I need the modification for the below mentioned code (found in one more post https://www.unix.com/shell-programming-scripting/27161-script-generate-average-values.html) to find the average values for all the columns(but for a specific rows) and print the averages side by side. I have...