读取.txt文件并将选择性数据导出到.csv

Question

I'm looking for help, I have multipath output from centos server in a .txt file, which looks like this.

asm (393040300403de) dm-12 HITACHI
size=35G queue_if_no_path
  |- 1:0:0:18  sda  65:48   active ready running
  `- 3:0:0:18  sdbc 70:368  active ready running
3600300300a4c dm-120 HITACHI
size=50G queue_if_no_path
  |- 1:0:0:98  sdc 70:48   active ready running
  `- 3:0:0:98  sdca 131:368 active ready running

It should look like this when exported to a .csv file.

DISKS_NAME  LUN             LUNID DM-NAME SIZE  MULTPATH
asm       393040300403de    03de  dm-12    35G  sda  sdbc
No_device  3600300300a4c    0a4c  dm-120   50G  sdc  sdca

This is as far i got, but this just reads every line and puts it into a different column every time it finds a space

import csv

readfile = 'multipath.txt'
writefile = 'data.csv'
with open(readfile,'r') as a, open(writefile, 'w') as b:
    o=csv.writer(b)
    for line in a:
        o.writerow(line.split())

Answer 1

Assuming that you only have the two types of entry as described in your above sample, you can define each line as a factor of the number of elements within it that will be seperated by line.split() . For example:

disk_name = ""
... # other parameters you need to keep track of across lines. I'd suggest creating a class for each lun/disk_name.

for line in a:
    line_data = line.split()

    if len(line_data) == 4:
        # this will match and 'asm (393040300403de) dm-12 HITACHI'
        disk_name, lun, dm_name, _ = line_data
        # process these variables accordingly (instantiate a new class member)
        continue # to read the next line

    else if len(line_data) == 3:
        # this will match '3600300300a4c dm-120 HITACHI'
        lun, dm_name, _ = line_data
        disk_name = "No_device"
        # process these variables accordingly
        continue

    if len(line_data) == 2:
        # this will match 'size=35G queue_if_no_path'
        size, _ = line_data
        # process the size accordingly, associate with the disk_name from earlier
        continue

    if len(line_data) == 7:
        # this will match '|- 1:0:0:18  sda  65:48   active ready running' etc.
        _, _, path, _, _, _, _ = line_data
        # process the path accordingly, associate with the disk_name from earlier
        continue

Of course, using a regex to work if the line contains the type of data that you need, rather than just the right number of items, will be more flexible. But this should get you started.

By processing the lines in this order, you'll always pick up a new disk_name / lun , and then assign the following "data" lines to that disk. When you hit a new disk, the lines following that will be associated with the new disk, etc.