Replace characters at a list of indices in a python string/list

Question

I have a dictionary as follows:

s_dict = {'s' : 'ATGCGTGACGTGA'}

I want to change the string stored as the value of the dictionary for key 's' at positions 4, 6, 7 and 10 to h, k, p and r.

pos_change = {'s' : ['4_h', '6_k', '7_p', '10_r']}

The way I can think about it is in a loop:

for key in s_dict:
 for position in pos_change[key]:
 pos = int(position.split("_")[0])
 char = position.split("_")[1]
 l = list(s_dict[key])
 l[pos]= char
 s_dict[key] = "".join(l)

Output:

s_dict = {'s': 'ATGChTkpCGrGA'}

This works fine but my actual s_dict file is about 1.5 Gb. Is there a faster way of replacing a list of characters at specific indices in a string or list?

Thanks!

• Chris_Rands: What are you doing to that DNA? Methylation?
• Ma0: pos_change would be better as a dict of dicts (pos_change = {'s' : {4: 'h', 6: 'k', 7: 'p', 10: 'r'}})
• Homap: @Chris_Rands Oh, no I just want to replace SNPs with IUPAC characters. I just made an example for the sake of the python question.
• user8060120: s_dict['s'] = '%s%s%s' % (s_dict['s'][:pos], char, s_dict['s'][pos+1:]) instead of do list and join
• Open AI - Opting Out: I'd use bytearray as it is mutable

Answer 1

Here is my take on your interesting problem:

s_dict = {'s' : 'ATGCGTGACGTGA'} 
pos_change = {'s' : ['4_h', '6_k', '7_p', '10_r']}

---

# 1rst change `pos_change` into something more easily usable
pos_change = {k: dict(x.split('_') for x in v) for k, v in pos_change.items()}
print(pos_change) # {'s': {'4': 'h', '6': 'k', '7': 'p', '10': 'r'}}

---

# and then... 
for k, v in pos_change.items():
 temp = set(map(int, v))
 s_dict[k] = ''.join([x if i not in temp else pos_change[k][str(i)] for i, x in enumerate(s_dict[k])])
print(s_dict) # {'s': 'ATGChTkpCGrGA'}

Answer 2

as an option of solution you can use s_dict['s'] = '%s%s%s' % (s_dict['s'][:pos], char, s_dict['s'][pos+1:]) instead of do list and join

In [1]: s_dict = {'s' : 'ATGCGTGACGTGA' * 10}
 ...: pos_change = {'s' : ['4_h', '6_k', '7_p', '10_r']}
 ...: 
 ...: def list_join():
 ...: for key in s_dict:
 ...: for position in pos_change[key]:
 ...: pos = int(position.split("_")[0])
 ...: char = position.split("_")[1]
 ...: l = list(s_dict[key])
 ...: l[pos]= char
 ...: s_dict[key] = "".join(l)
 ...: 
 ...: def by_str():
 ...: for key in s_dict:
 ...: for position in pos_change[key]:
 ...: pos = int(position.split("_")[0])
 ...: char = position.split("_")[1]
 ...: values = s_dict['s'][:pos], char, s_dict['s'][pos+1:]
 ...: s_dict['s'] = '%s%s%s' % values
 ...: 
In [2]: %timeit list_join()
11.7 μs ± 191 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each)
In [3]: %timeit by_str()
4.29 μs ± 46.7 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each)