Python：如何遍历行块并在行内复制特定文本

Question

Input file:

DATE: 07/01/15 @ 0800                 HYRULE HOSPITAL                         PAGE 1
USER: LINK                      Antibiotic Resistance Report
--------------------------------------------------------------------------------------------
Activity Date Range: 01/01/15 - 02/01/15
--------------------------------------------------------------------------------------------
HH0000000001 LINK,DARK 30/M <DIS IN 01/05> (UJ00000001) A001-01 0A ZELDA,PRINCESS MD
15:M0000001R    COMP, Coll: 01/02/15-0800 Recd: 01/02/15-0850 (R#00000001) ZELDA,PRINCESS MD
    Source: SPUTUM                                  
       PSEUDOMONAS FLUORESCENS            LEVOFLOXACIN   >=8   R                            
--------------------------------------------------------------------------------------------
HH0000000002 FAIRY,GREAT   25/F <DIS IN 01/06> (UJ00000002) A002-01 0A ZELDA,PRINCESS MD    
15:M0000002R    COMP, Coll: 01/03/15-2025 Recd: 01/03/15-2035 (R#00000002) ZELDA,PRINCESS MD
    Source: URINE- STRAIGHT CATH                    
   PROTEUS MIRABILIS                  CEFTRIAXONE-other      R                          
--------------------------------------------------------------------------------------------
HH0000000003 MAN,OLD   85/M <DIS IN 01/07> (UJ00000003) A003-01 0A ZELDA,PRINCESS MD 
15:M0000003R    COMP, Coll: 01/04/15-1800 Recd: 01/04/15-1800 (R#00000003) ZELDA,PRINCESS MD
    Source: URINE-CLEAN VOIDED SPEC                 
   ESCHERICHIA COLI                   LEVOFLOXACIN   >=8   R                            
--------------------------------------------------------------------------------------------

Completely new to programming/scripting and Python. How do you recommend looping through this sample input to grab specific text in the fields?

Each patient has a unique identifier (eg HH0000000001). I want to grab specific text from each line.

Output should look like:

Date|Time|Name|Account|Specimen|Source|Antibiotic
01/02/15|0800|LINK, DARK|HH0000000001|PSEUDOMONAS FLUORESCENS|SPUTUM|LEVOFLOXACIN
01/03/15|2025|FAIRY, GREAT|HH0000000002|PROTEUS MIRABILIS|URINE- STRAIGHT CATH|CEFTRIAXONE-other

Edit: My current code looks like this:

(Disclaimer: I am fumbling around in the dark, so the code is not going to be pretty at all.

input = open('report.txt')
output = open('abx.txt', 'w')

date = ''  # Defining global variables outside of the loop
time = ''
name = ''
name_last = ''
name_first = ''
account = ''
specimen = ''
source = ''

output.write('Date|Time|Name|Account|Specimen|Source\n')
lines = input.readlines()

for index, line in enumerate(lines):
    print index, line

    if last_line_location:
        new_patient = True
        if not first_time_through:
            output.write("{}|{}|{}, {}|{}|{}|{}\n".format(
                'Date', # temporary placeholder
                'Time', # temporary placeholder
                name_last.capitalize(),
                name_first.capitalize(),
                account,
                'Specimen', # temporary placeholder
                'Source' # temporary placeholder
                ) )
        last_line_location = False
        first_time_through = False

    for each in lines:
        if line.startswith('HH'):  # Extract account and name
            account = line.split()[0]
            name = line.split()[1]
            name_last = name.split(',')[0]
            name_first = name.split(',')[1]
            last_line_location = True

input.close()
output.close()

Currently, the output will skip the first patient and will only display information for the 2nd and 3rd patient. Output looks like this:

Date|Time|Name|Account|Specimen|Source
Date|Time|Fairy, Great|HH0000000002|Specimen|Source
Date|Time|Man, Old|HH0000000003|Specimen|Source

Please feel free to make suggestions on how to improve any aspect of this, including output style or overall strategy.

Answer 1

You code actually works if you add...

last_line_location = True
first_time_through = True

...before your for loop

You asked for pointers as well though...

As has been suggested in the comments, you could look at the re module.

I've knocked something together that shows this. It may not be suitable for all data because three records is a very small sample, and I've made some assumptions.
The last item is also quite contrived because there's nothing definite to search for (such as Coll , Source ). It will fail if there are no spaces at the start of the final line, for example.

This code is merely a suggestion of another way of doing things:

import re

startflag = False
with open('report.txt','r') as infile:
    with open('abx.txt','w') as outfile:
        outfile.write('Date|Time|Name|Account|Specimen|Source|Antibiotic\n')
        for line in infile:
            if '---------------' in line:
                if startflag:
                    outfile.write('|'.join((date, time, name, account, spec, source, anti))+'\n')
                else:
                    startflag = True
                continue
            if 'Activity' in line:
                startflag = False

            acc_name = re.findall('HH\d+ \w+,\w+', line)
            if acc_name:
                account, name = acc_name[0].split(' ')

            date_time = re.findall('(?<=Coll: ).+(?= Recd:)', line)
            if date_time:
                date, time = date_time[0].split('-')

            source_re = re.findall('(?<=Source: ).+',line)
            if source_re:
                source = source_re[0].strip()

            anti_spec = re.findall('^ +(?!Source)\w+ *\w+ + \S+', line)
            if anti_spec:
                stripped_list = anti_spec[0].strip().split()
                anti = stripped_list[-1]
                spec = ' '.join(stripped_list[:-1])

Output

Date|Time|Name|Account|Specimen|Source|Antibiotic
01/02/15|0800|LINK,DARK|HH0000000001|PSEUDOMONAS FLUORESCENS|SPUTUM|LEVOFLOXACIN
01/03/15|2025|FAIRY,GREAT|HH0000000002|PROTEUS MIRABILIS|URINE- STRAIGHT CATH|CEFTRIAXONE-other
01/04/15|1800|MAN,OLD|HH0000000003|ESCHERICHIA COLI|URINE-CLEAN VOIDED SPEC|LEVOFLOXACIN

Edit:
Obviously, the variables should be reset to some dummy value between writes on case of a corrupt record. Also, if there is no line of dashes after the last record it won't get written as it stands.