I know similar questions has been answered many time on SO.
(one example is here: https://stackoverflow.com/questions/7846476/replace-column-in-one-file-with-column-from-another-using-awk)
But, this is something unique in my case as I need to take care of a particular pattern.
header of my file1, that I want to get updated, is
3 6 0 6.0361821 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000
0.994429353 0.000000000 0.000000000
0.000000000 0.994429353 0.000000000
0.000000000 0.000000000 2.469627493
1 'A ' 63548.626894188397
2 'B ' 169717.29799472401
3 'C ' 25598.367262405900
1 2 0.7458220147 0.7458220147 1.8031927376 << need to be updated from here
2 2 0.2486073382 0.2486073382 0.6664347554
3 1 0.2486073382 0.2486073382 2.2628589536
4 1 0.7458220147 0.7458220147 0.2067685394
5 3 0.7458220147 0.7458220147 1.0275486366
6 3 0.2486073382 0.2486073382 1.4420788564 << upto here
T
21.3496599 0.0000000 0.0000000
0.0000000 21.3496599 0.0000000
0.0000000 0.0000000 24.1101752
1
-7.6119990 -0.0000000 0.0000000
0.0000000 -7.6119990 0.0000000
0.0000000 0.0000000 -7.0331945
2
-7.6119990 0.0000000 0.0000000
-0.0000000 -7.6119990 0.0000000
0.0000000 0.0000000 -7.0331945
3
3.4711749 0.0000000 0.0000000
I need to update the $2, $3 and $4 of file1 from $ESPi"th line to "$ESPf"th with $1, $2 and $3 of file2 (mentioned below). The spaces in file1 should not change while updating. Here $ESPi"th and "$ESPf"th represents 8th line and 13th lines, respectively and changes case to case.
file2 is
0.750000000 0.750000000 0.730147661 << with these data
0.250000000 0.250000000 0.269852339
0.250000000 0.250000000 0.916275414
0.750000000 0.750000000 0.083724586
0.750000000 0.750000000 0.416074343
0.250000000 0.250000000 0.583925657 < upto these data
I have tried to do my job with.
#!/bin/bash
for j in `seq "$ESPi" 1 "$ESPf"` # ESPi and ESPf are 8 and 13, respectively here and change case by case.
do
ESP1=$(cat file1 | head -n "$j" | tail -n 1 | awk '{print $3}')
ESP2=$(cat file1 | head -n "$j" | tail -n 1 | awk '{print $4}')
ESP3=$(cat file1 | head -n "$j" | tail -n 1 | awk '{print $5}')
for k in `seq 1 1 "$NELEMENTS"` # $NELEMENTS is six here.
do
qeIN1=$(cat file2 | head -n "$k" | tail -n 1 | awk '{print $1}')
qeIN2=$(cat file2 | head -n "$k" | tail -n 1 | awk '{print $2}')
qeIN3=$(cat file2 | head -n "$k" | tail -n 1 | awk '{print $3}')
sed 's/'$ESP1'/'$qeIN1'/g' file1
sed 's/'$ESP2'/'$qeIN2'/g' file1
sed 's/'$ESP3'/'$qeIN3'/g' file1
done
done
This gives me
3 6 0 6.0361821 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000
0.994429353 0.000000000 0.000000000
0.000000000 0.994429353 0.000000000
0.000000000 0.000000000 2.469627493
1 'A ' 63548.626894188397
2 'B ' 169717.29799472401
3 'C ' 25598.367262405900
1 2 0.7458220147 0.7458220147 1.8031927376
2 2 0.750000000 0.750000000 0.6664347554
3 1 0.750000000 0.750000000 2.2628589536
4 1 0.7458220147 0.7458220147 0.2067685394
5 3 0.7458220147 0.7458220147 1.0275486366
6 3 0.750000000 0.750000000 1.4420788564
T
21.3496599 0.0000000 0.0000000
0.0000000 21.3496599 0.0000000
0.0000000 0.0000000 24.1101752
1
-7.6119990 -0.0000000 0.0000000
0.0000000 -7.6119990 0.0000000
0.0000000 0.0000000 -7.0331945
2
-7.6119990 0.0000000 0.0000000
-0.0000000 -7.6119990 0.0000000
0.0000000 0.0000000 -7.0331945
3
3.4711749 0.0000000 0.0000000
The expected output is
3 6 0 6.0361821 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000
0.994429353 0.000000000 0.000000000
0.000000000 0.994429353 0.000000000
0.000000000 0.000000000 2.469627493
1 'A ' 63548.626894188397
2 'B ' 169717.29799472401
3 'C ' 25598.367262405900
1 2 0.750000000 0.750000000 0.730147661
2 2 0.250000000 0.250000000 0.269852339
3 1 0.250000000 0.250000000 0.916275414
4 1 0.750000000 0.750000000 0.083724586
5 3 0.750000000 0.750000000 0.416074343
6 3 0.250000000 0.250000000 0.583925657
T
21.3496599 0.0000000 0.0000000
0.0000000 21.3496599 0.0000000
0.0000000 0.0000000 24.1101752
1
-7.6119990 -0.0000000 0.0000000
0.0000000 -7.6119990 0.0000000
0.0000000 0.0000000 -7.0331945
2
-7.6119990 0.0000000 0.0000000
-0.0000000 -7.6119990 0.0000000
0.0000000 0.0000000 -7.0331945
3
3.4711749 0.0000000 0.0000000
I am looking for shell (bash) script.
#!/bin/bash
ESPi=8
ESPf=13
python > file1.new <<EOF
import sys, re
write = sys.stdout.write
espi = $ESPi
espf = $ESPf
repls = {2:0, 3:1, 4:2}
with open("file1") as f1, open("file2") as f2:
for i in range(espi - 1): write(next(f1))
for i in range(espf - espi + 1):
line = next(f1)
toks = next(f2).split()
for col, rcol in repls.items():
pat = "(\s*)((\S+\s+){{{col}}})(\S+)(.*)".format(col=col)
repl = r"\g<1>\g<2>{val}\g<5>".format(val=toks[rcol])
line = re.sub(pat, repl, line)
write(line)
for line in f1: write(line)
EOF
mv file1.new file1