Question

我想删除CSV文件字段数据中的新行字符。 SO /其他地方的多个人也会问同样的问题。但是，提供的解决方案是脚本编写的。我正在寻找像PYTHON这样的编程语言或Spark中的解决方案（不仅仅是这两个），因为我有很大的文件。

以前就同一主题提出的问题：

我有一个大小约为1GB的CSV文件，想要删除字段数据中的新行字符。 CSV文件的架构动态变化，因此我无法对架构进行硬编码。换行符并不总是出现在逗号之前，即使在字段中也会随机出现。

示例数据：

playerID,yearID,gameNum,gameName,teamName,lgID,GP,startingPos
gomezle01,1933,1,Cricket,Team1,NYA,AL,1
ferreri01,1933,2,Hockey,"This is 
Team2",BOS,AL,1
gehrilo01,1933,3,"Game name is 
Cricket" 
,Team3,NYA,AL,1
gehrich01,1933,4,Hockey,"Here it is 
Team4",DET,AL,1
dykesji01,1933,5,"Game name is 
Hockey"
,"Team name 
Team5",CHA,AL,1

预期输出：

playerID,yearID,gameNum,gameName,teamName,lgID,GP,startingPos
gomezle01,1933,1,Cricket,Team1,NYA,AL,1
ferreri01,1933,2,Hockey,"This is Team2",BOS,AL,1
gehrilo01,1933,3,"Game name is Cricket" ,Team3,NYA,AL,1
gehrich01,1933,4,Hockey,"Here it is Team4",DET,AL,1
dykesji01,1933,5,"Game name is Hockey","Team name Team5",CHA,AL,1

换行符可以包含在任何字段的数据中。

修改根据代码截图：

Answer 1

您可以使用re，pandas和io模块，如下所示：

import re
import io
import pandas as pd

with open('data.csv','r') as f:
    data = f.read()
df = pd.read_csv(io.StringIO(re.sub('"\s*\n','"',data)))

for col in df.columns: #To replace all line breaks in all textual columns
    if df[col].dtype == np.object_:
        df[col] = df[col].str.replace('\n','');

In [78]: df
Out[78]:
    playerID    yearID  gameNum gameName               teamName        lgID GP  startingPos
0   gomezle01   1933    1       Cricket                Team1           NYA  AL  1
1   ferreri01   1933    2       Hockey                 This is Team2   BOS  AL  1
2   gehrilo01   1933    3       Game name is Cricket   Team3           NYA  AL  1
3   gehrich01   1933    4       Hockey  Here it is     Team4           DET  AL  1
4   dykesji01   1933    5       Game name is Hockey    Team name Team5 CHA  AL  1

如果您希望此DataFrame作为输出CSV文件使用：

df.to_csv('./output.csv')

Answer 2

这是一个基本的，在通过csv读取之前进行简单的预处理。

import csv

def simple_sanitize(data):
    result = []
    for i, a in enumerate(data):
        if i + 1 != len(data) and data[i + 1][0] == ',':
            a = a.replace('\n', '')
            result.append(a + data[i + 1])
        elif a[0] != ',':
            result.append(a)
    return result

data = [line for line in open('test.csv', 'r')]
sdata = simple_sanitize(data)

with open('out.csv','w') as f:
    for row in sdata:
        f.write(row)

result = [list(val.replace('\n', '') for val in line) for line in csv.reader(open('out.csv', 'r'))]

print(result)

结果：

[['playerID', 'yearID', 'gameNum', 'gameName', 'teamName', 'lgID', 'GP', 'startingPos'], 
['gomezle01', '1933', '1', 'Cricket', 'Team1', 'NYA', 'AL', '1'], 
['ferreri01', '1933', '2', 'Hockey', 'This is Team2', 'BOS', 'AL', '1'], 
['gehrilo01', '1933', '3', 'Game name is Cricket ', 'Team3', 'NYA', 'AL', '1'], 
['gehrich01', '1933', '4', 'Hockey', 'Here it is Team4', 'DET', 'AL', '1'], 
['dykesji01', '1933', '5', 'Game name is Hockey', 'Team name Team5', 'CHA', 'AL', '1']]

Answer 3

它可以使用一点清洁，但这里有一些代码可以做你想要的。适用于字段内和逗号之前的换行符。如果需要更多要求，可以进行一些调整：

import csv

with open('data.csv', 'r') as csvfile:
    reader = csv.reader(csvfile, delimiter=',', quotechar='"')
    actual_rows = [next(reader)]
    length = len(actual_rows[0])
    real_row = []
    for row in reader:
        if len(row) < length:
            if real_row:
                real_row[-1] += row[0]
                real_row += row[1:]
            else:
                real_row = row
        else:
            real_row = row
        if len(real_row) == length:
            real_row = map(lambda s: s.replace('\n', ' '), real_row)
            # store real_row or use it as needed
            actual_rows.append(list(real_row))
            real_row = []

    print(actual_rows)

我将修正后的行存储在actual_rows中，但如果您不想加载到内存中，只需在评论中指出的每个循环中使用real_row变量

Answer 4

此解决方案的基本思想是使用grouper recipe获取固定长度的块（长度等于第一行中的列数）。由于它不会立即读取整个文件，因此不会因大文件而耗尽内存使用量。

$ cat a.py
import csv,itertools as it,operator as op

def grouper(iterable,n):return it.zip_longest(*[iter(iterable)]*n)

with open('in.csv') as inf,open('out.csv','w',newline='') as outf:
 r,w=csv.reader(inf),csv.writer(outf)
 hdr=next(r)
 w.writerow(hdr)
 for row in grouper(filter(bool,map(op.methodcaller('replace','\n',''),it.chain.from_iterable(r))),len(hdr)):
  w.writerow(row)

$ python3 a.py
$ cat out.csv
playerID,yearID,gameNum,gameName,teamName,lgID,GP,startingPos
gomezle01,1933,1,Cricket,Team1,NYA,AL,1
ferreri01,1933,2,Hockey,This is Team2,BOS,AL,1
gehrilo01,1933,3,Game name is Cricket ,Team3,NYA,AL,1
gehrich01,1933,4,Hockey,Here it is Team4,DET,AL,1
dykesji01,1933,5,Game name is Hockey,Team name Team5,CHA,AL,1

这里假设的一个假设是输入csv中没有空单元格。

从CSV文件中删除新行

4 个答案: