将txt文件解析为字典以写入csv文件

Question

Eprime输出一个.txt文件,如下所示:

*** Header Start ***
VersionPersist: 1
LevelName: Session
Subject: 7
Session: 1
RandomSeed: -1983293234
Group: 1
Display.RefreshRate: 59.654
*** Header End ***
    Level: 2
    *** LogFrame Start ***
    MeansEffectBias: 7
    Procedure: trialProc
    itemID: 7
    bias1Answer: 1
    *** LogFrame End ***
    Level: 2
    *** LogFrame Start ***
    MeansEffectBias: 2
    Procedure: trialProc
    itemID: 2
    bias1Answer: 0

我想解析它并将其写入.csv文件,但删除了多行.

我试图创建一个字典,将冒号前面的文本作为键,然后将文本作为值:

 {subject: [7, 7], bias1Answer : [1, 0], itemID: [7, 2]} 

def load_data(filename):
    data = {}
    eprime = open(filename, 'r')
    for line in eprime:
        rows = re.sub('\s+', ' ', line).strip().split(':')
        try:
            data[rows[0]] += rows[1]
        except KeyError:
            data[rows[0]] = rows[1]
    eprime.close()
    return data

for line in open(fileName, 'r'):
    if ':' in line:
        row = line.strip().split(':')
        fullDict[row[0]] = row[1]
print fullDict

以下两个脚本都会产生垃圾:

{'\x00\t\x00M\x00e\x00a\x00n\x00s\x00E\x00f\x00f\x00e\x00c\x00t\x00B\x00i\x00a\x00s\x00': '\x00 \x005\x00\r\x00', '\x00\t\x00B\x00i\x00a\x00s\x002\x00Q\x00.\x00D\x00u\x00r\x00a\x00t\x00i\x00o\x00n\x00E\x00r\x00r\x00o\x00r\x00': '\x00 \x00-\x009\x009\x009\x009\x009\x009\x00\r\x00'

如果我可以设置字典,我可以将它写入一个看起来像这样的csv文件!!:

 Subject  itemID ... bias1Answer 
  7       7             1
  7       2             0

Answer 1

您不需要创建字典.

import codecs
import csv

with codecs.open('eprime.txt', encoding='utf-16') as f, open('output.csv', 'w') as fout:
    writer = csv.writer(fout, delimiter='\t')
    writer.writerow(['Subject', 'itemID', 'bias1Answer'])
    for line in f:
        if ':' in line:
            value = line.split()[-1]

        if 'Subject:' in line:
            subject = value
        elif 'itemID:' in line:
            itemID = value
        elif 'bias1Answer:' in line:
            bias1Answer = value
            writer.writerow([subject, itemID, bias1Answer])