Bash：如何从CSV文件的列中抓取带有“定界符”的列？

Question

I have a few CSV files I have downloaded from an online database. I am trying to cut them down so that I can insert the portions of the data that I need into my SQL database.

The CSV file has comma-separated fields and looks like this:

1,Peptidoglycan synthetase ftsI,ftsI,1574687,L42023,P45059,FTSI_HAEIN,"",,,,
3,Histidine decarboxylase,HDC,32109,X54297,P19113,DCHS_HUMAN,,HDC,HDC,HGNC:4855,00817
5,"Glutaminase liver isoform, mitochondrial",GLS2,6650606,AF110330,Q9UI32,GLSL_HUMAN,,GLS2,GLS2,HGNC:29570,05901
6,Coagulation factor XIII A chain,F13A1,182309,M22001,P00488,F13A_HUMAN,1FIE,F13A1,F13A1,HGNC:3531,00604
7,"Nitric oxide synthase, inducible",NOS2,292242,L09210,P35228,NOS2_HUMAN,2NSI,NOS2A,NOS2A,HGNC:7873,01225

And here is the problem. Look at the 3rd and the 5th lines. The 2nd column of those two lines have commas in them! I usually use awk for something like this, and because of that comma being there $2 gets messed up.

So for example:

awk -F ',' '{print $2}' myfile.csv ## Obviously I will be printing a lot more stuff

If that was done to that portion as shown above, the 3rd and 5th line will be screwed up because 2nd column of those two lines have a comma in them enclosed by quotation marks.

What could I do to get around this?

EDIT: I'd like to still stick to doing this in shell if possible.

Answer 1

You should use a CSV parser like perl 's Text::CSV (in a one-liner (so still in shell) if you want), it will do all the magic for you.

If instead you prefer python , see the csv module

An example in shell + python :

$ python<<EOF
import csv

f = open("test.csv", 'rt')
try:
    reader = csv.reader(f)
    for row in reader:
        print row
finally:
    f.close()
EOF