data transformation python,使用正则表达式将数据转换为Python中的字典

I have a dataset with FASTA formatted sequencing, basically like this:

>pc284

ATCGCGACTCGAC

>pc293

ACCCGACCTCAGC

I want to take to use each tag as a key in the dictionary, and store the gene as a value.

This is the code I have, but really isn't doing anything:

import re

fileData = open('d.fasta', 'r')

myDict = dict()

for line in fileData:

match = re.search('(\>)(\w+)(\r)(\w+)', line)

if match:

gene = match.group(3)

myDict[gene[0]] = gene[1]

print myDict

解决方案

\r is not a valid character class, I think you meant to use \s instead. You can reduce the groups if you don't use them either.

But most of all, you need to extract your groups correctly:

match = re.search(r'>(\w+)\s+(\w+)', line)

if match:

tag, gene = match.groups()

myDict[tag] = gene

By creating only two capturing groups, we can more simply extract those two with .groups() and directly assign them to two variables, tag and gene.

However, reading up on the FASTA format seems to indicate this is a multi-line format with the tag on one line, the gene data on multiple lines after that. In that case your \r was meant to match the newline. This won't work as you read the file one line at a time.

It would be much simpler to read that format without regular expressions like so:

myDict = {}

with open('d.fasta', 'rU') as fileData:

tag = None

for line in fileData:

line = line.strip()

if not line:

continue

if line[0] == '>':

tag = line[1:]

myDict[tag] = ''

else:

assert tag is not None, 'Invalid format, found gene without tag'

myDict[tag] += line

print myDict

This reads the file line by line, detecting tags based on the starting > character, then reads multiple lines of gene information collecting it into your dictionary under the most-recently read tag.

Note the rU mode; we open the file using python's universal newlines mode, to handle whatever newline convention was used to create the file.

Last but not least; take a look at the BioPy project; their Bio.SeqIO module handles FASTA plus many other formats perfectly.