在python中过滤CSV文件

Question

我已经下载了这个csv file，它创建了一个基因信息电子表格。重要的是在HLA-* 列中，有基因信息。如果基因的分辨率太低，例如DQB1*03 那么该行应该被删除。如果数据分辨率太高，例如DQB1*03:02:01，那么末尾的:01标签需要去掉。因此，理想情况下，我希望蛋白质采用DQB1*03:02 格式，以便在DQB1* 之后具有两个级别的分辨率。我如何告诉 python 寻找这些格式，并忽略其中存储的数据。 例如

if (csvCell is of format DQB1*03:02:01):
   delete the :01 # but do this in a general format
elif (csvCell is of format DQB1*03):
   delete row
else:
   goto next line

更新：我引用的编辑代码

import csv
import re
import sys

csvdictreader = csv.DictReader(open('mhc.csv','r+b'), delimiter=',')
csvdictwriter = csv.DictWriter(file('mhc_fixed.csv','r+b'), fieldnames=csvdictreader.fieldnames, delimiter=',')
csvdictwriter.writeheader()
targets = [name for name in csvdictreader.fieldnames if name.startswith('HLA-D')]

for rowfields in csvdictreader:
  keep = True
  for field in targets:
    value = rowfields[field]
    if re.match(r'^\w+\*\d\d$', value):
      keep = False
      break # quit processing target fields
    elif re.match(r'^(\w+)\*(\d+):(\d+):(\d+):(\d+)$', value):
      rowfields[field] = re.sub(r'^(\w+)\*(\d+):(\d+):(\d+):(\d+)$',r'\1*\2:\3', value)
    else: # reduce gene resolution if too high
              # by only keeping first two alles if three are present
      rowfields[field] = re.sub(r'^(\w+)\*(\d+):(\d+):(\d+)$',r'\1*\2:\3', value)
  if keep:
     csvdictwriter.writerow(rowfields)

Answer 1

这是我认为可以满足您的需求的东西。它不像彼得的回答那么简单，因为它使用 Python 的 csv 模块来处理文件。它可能会被重写和简化，就像他一样将文件视为纯文本，但这应该很容易。

import csv
import re
import sys

csvdictreader = csv.DictReader(sys.stdin, delimiter=',')
csvdictwriter = csv.DictWriter(sys.stdout, fieldnames=csvdictreader.fieldnames, delimiter=',')
csvdictwriter.writeheader()
targets = [name for name in csvdictreader.fieldnames if name.startswith('HLA-')]

for rowfields in csvdictreader:
    keep = True
    for field in targets:
        value = rowfields[field]
        if re.match(r'^DQB1\*\d\d$', value): # gene resolution too low?
            keep = False
            break # quit processing target fields
        else: # reduce gene resolution if too high
              # by only keeping first two alles if three are present
            rowfields[field] = re.sub(r'^DQB1\*(\d\d):(\d\d):(\d\d)$',
                                      r'DQB1*\1:\2', value)
    if keep:
        csvdictwriter.writerow(rowfields)

对我来说最困难的部分是确定你想做什么。

Answer 2

这是一个超简单的过滤器：

import sys

for line in sys.stdin:
  line = line.replace( ',DQB1*03:02:01,', ',DQB1*03:02,' )

  if line.find( ',DQB1*03,' ) == -1:
    sys.stdout.write( line )

或者，如果你想使用正则表达式

import re
import sys

for line in sys.stdin:
  line = re.sub( ',DQB1\\*03:02:01,', ',DQB1*03:02,', line )
  if re.search( ',DQB1\\*03,', line ) == None:
    sys.stdout.write( line )

运行它

python script.py < data.csv