在bash中按特定顺序打印文件

Question

I have file with content:

file.txt:

Iteration 1
RAM: +456ms
Cache: +142ms (total +417ms)

Iteration 2
Spec: +152ms
Cache: +149ms (total +413ms)

Iteration 3
RAM: +184ms
Spec: +172ms
Searchms: +131ms (total +385ms)

First launch 4
RAM: +149ms
Searchms: +188ms

In this file between every First launch , content can be different, it is not fixed ( for example: First launch 3 contain three elements while First launch 2 contents only 2 elements), so any number of content can be between the First launch pattern which is not known at the beginning.

Expected Output:

RAM  456 184 149
Cache 142 149  
Spec 152 172
Searchms 131 188

Due to not knowing exact approach,I have tried this code.

My code:

for i in {1..4}
do
awk "/First launch $i/{flag=1;next} /First launch $((i+1))/{flag=0} flag" file.txt> fl$i.txt
sed -i 's/\+//g' fl$i.txt
sed -i 's/://g' fl$i.txt
sed -i 's/(.*//g' fl$i.txt
sed -i 's/ms//g' fl$i.txt
awk '{print $1 "\t" $2}' fl$i.txt
done

My output has two issues: I am generating the file for each pattern which is wrong. Also I wanted to remove ms after the time but it also removes the ms from the pattern name (ex: Searchms to Search)

Output:

fl1.txt: 
    RAM     456
    Cache   142
fl2.txt : 
    Spec    152
    Cache   149
fl3.txt  :
    RAM     184
    Spec    152
    Search  131
fl4.txt : 
    RAM     149
    Search  188

Please suggest me an approach to get the expected output without generating any extra file with the constraint of removing ms after the time.

Answer 1

One using awk:

$ awk '
$1 !~ /^(|First)$/ {            # avoid forbidden keywords and empty lines
    gsub(/[^0-9]/,"",$2)        # remove non-numerals
    a[$1]=a[$1] OFS $2          # append to associative array 
}
END {                           # in the end
    for(i in a)                 # loop all keywords
        print i a[i]            # output
}' file

Output lines in awk default order (appears random):

Cache: 142 149
Searchms: 131 188
Spec: 152 172
RAM: 456 184 149

Answer 2

$ cat tst.awk
BEGIN { FS="[: ]+" }
/:/ { vals[$1] = vals[$1] OFS $2+0 }
END { for (key in vals) print key vals[key] }

$ awk -f tst.awk file
Cache 142 149
RAM 456 184 149
Searchms 131 188
Spec 152 172

Answer 3

Perl one-liner:

$ perl -nE 'if (/^(\w+):\s+\+(\d+)ms/) { push @{$keys{$1}}, $2 } END { while (($k, $vs) = each %keys) { say join(" ", $k, @$vs) }}' file.txt
Spec 152 172
Searchms 131 188
Cache 142 149
RAM 456 184 149

(Order of the lines will vary; pipe it to sort if that matters)

---

How it works:

For each line in the file, if it matches the regular expression ^(\\w+):\\s+\\+(\\d)ms (1 or more alphanumeric characters at the start of the line, followed by a colon, whitespace, a plus sign, 1 or more digits, and then the letters m and s), it adds the number to the appropriate array in a hash using the starting word as the key. Then it prints out all those starting words and their associated arrays.

Basically the same idea as the awk answer, but that uses strings instead of arrays because awk doesn't have true arrays like perl, just associative ones (Which are called hashes in perl lingo).

Answer 4

This might work for you (GNU sed):

sed -E '/:/!d;s/.([0-9]+).*/\1/;H;x;s/((\n\S+) \S+)(.*)\2(.*)/\1\4\3/;h;$!d;x;s/.//' file

Any lines other than those that contain a : are noise, so delete them.

Remove all but the key, a space and the first set of integers from each line.

Append the result to the hold space.

Using pattern matching, gather up like keys data and retain in the hold space.

Delete all but the last line.

At the end of file, swap to the hold space, remove the introduced newline and print the result.

Answer 5

Another Perl, using paragraph mode -00

perl -00 -lnE ' 
               while(/(^\S+):.+?(\d+)/gm ) {push(@{$kv{$1}},$2)} 
            END { foreach(keys %kv) { print "$_ @{$kv{$_}}" } }    '

with inputs

$ cat arya.txt
First launch 1
RAM: +456ms
Cache: +142ms (total +417ms)

First launch 2
Spec: +152ms
Cache: +149ms (total +413ms)

First launch 3
RAM: +184ms
Spec: +172ms
Searchms: +131ms (total +385ms)

First launch 4
RAM: +149ms
Searchms: +188ms

$ perl -00 -lnE ' while(/(^\S+):.+?(\d+)/gm ) {push(@{$kv{$1}},$2)} END { foreach(keys %kv) { print "$_ @{$kv{$_}}" } } ' arya.txt
RAM 456 184 149

Cache 142 149

Searchms 131 188

Spec 152 172


$

Answer 6

Simply:

#!/bin/bash
for i in RAM Cache Spec Searchms; do
    echo "$i `cat file.txt  | grep $i | grep -Eo '[0-9]{1,9}' | tr '\n' ' '`" >> out.txt
done

You can change the order in line 2 ( for loop )

Output:

$ cat out.txt 
RAM 456 184 149 
Cache 142 417 149 413 
Spec 152 172 
Searchms 131 385 188