如何使用 Python 计算嵌套字典 in.json 文件中某个项目的出现次数（并可能迭代）？

Question

I'm trying for some time but since I'm still quite a begginer I'm having a hard time. I have a file with jsons and all of them have this structure:

{
   "cds":{
      "ENSLAFT00000035968.1":{
         "A":407,
         "C":312,
         "G":320,
         "T":320,
         "Y":0,
         "M":0,
         "S":0,
         "R":0,
         "W":0,
         "K":0,
         "N":0,
         "D":0,
         "B":0,
         "H":0,
         "V":0,
         "all":1359
      },
      "cdna":{
         "ENSLAFT00000034174.1":{
            "A":825,
            "C":700,
            "G":663,
            "T":584,
            "Y":0,
            "M":0,
            "S":0,
            "R":0,
            "W":0,
            "K":0,
            "N":0,
            "D":0,
            "B":0,
            "H":0,
            "V":0,
            "all":2772
         }
      }
   }

The first keys (cds and cdna) have each about over 1000 values (genes, the ENSLAFT+number). I would like to count all of the "N" occurrences (if some has fe 50 and some has 10, add them together and have 60). Shall I use Counter from collections or sum() or len() or some combination of them somehow...? And how to make a cycle like that for each file in my folder with jsons with the same structure? It sounds easy for me but I don't have much experience, so far I'm only able to count using pandas DataFrame or with not so complicated data...

I appreciate any help and further study recommendations!

Answer 1

You can use a for-loop:

data = {"cds": {"ENSLAFT00000035968.1": {"A": 407, "C": 312, "G": 320, "T": 320, "Y": 0, "M": 0, "S": 0, "R": 0, "W": 0, "K": 0, "N": 0, "D": 0, "B": 0, "H": 0, "V": 0, "all": 1359}}, "cdna": {"ENSLAFT00000034174.1": {"A": 825, "C": 700, "G": 663, "T": 584, "Y": 0, "M": 0, "S": 0, "R": 0, "W": 0, "K": 0, "N": 0, "D": 0, "B": 0, "H": 0, "V": 0, "all": 2772}}}
counter = 0
for value in data.values():
    # key would be cds or cdna, value is the dict of genes
    for gene in value.values():
        # key would be ENSLAFT00000035968.1, ...
        if 'N' in gene:
            counter += gene['N']
print(counter)

You can check the key to only count some:

counter = 0
for key, value in data.items():
    # key would be cds or cdna, value is the dict of genes
    if key == "cds":
        for gene in value.values():
            # key would be ENSLAFT00000035968.1, ...
            if 'N' in gene:
                counter += gene['N']
print(counter)

Answer 2

You could go about it by brute force looking for a regex in a string version of your JSON, eg

import json
import re

s = {"cds": {"ENSLAFT00000035968.1": {"A": 407, "C": 312, "G": 320, "T": 320, "Y": 0, "M": 0, "S": 0, "R": 0, "W": 0, "K": 0, "N": 0, "D": 0, "B": 0, "H": 0, "V": 0, "all": 1359}, "cdna": {"ENSLAFT00000034174.1": {"A": 825, "C": 700, "G": 663, "T": 584, "Y": 0, "M": 0, "S": 0, "R": 0, "W": 0, "K": 0, "N": 0, "D": 0, "B": 0, "H": 0, "V": 0, "all": 2772}}}}


s_str = json.dumps(s)
m = re.findall(r'"N":\s(\d+)', s_str)
print(m)  # prints ['0', '0']
print(len(m))  # prints 2

Or go the cleaner, longer route of a recursive function...

def rec_find(s, cur=0):
  if type(s) not in (dict, ):
    return 0
  resp = 0
  if "N" in s.keys():
    resp += 1
  for k in s.keys():
    resp += rec_find(s[k], resp)
  return resp

print(rec_find(s))  # prints 2