从头到尾计算 CSV 行中的沿袭

Question

下面我有一个 CSV 文件，其中每一列都包含一个谱系。每列都有不同长度的谱系。我试着从谱系的末尾开始计数，就像我从最后一个元素到谱系的开始计数一样。

Column1             Column2             Column3
root                root                root
cellular organisms  cellular organisms  cellular organisms
Eukaryota           Eukaryota           Eukaryota
Sar                 Sar                 Viridiplantae
Alveolata           Alveolata          
Apicomplexa         Apicomplexa
Aconoidasida

我尝试了@xjcl 提供的以下代码，但问题是脚本认为所有谱系行的长度都相同，因此产生了错误的值。任何帮助

import pandas as pd
from io import StringIO

def filter_and_count(df, search_string):
    df_filtered = df.loc[:, (df == search_string).any(axis=0)] # to access a group of rows and columns by label
    return pd.melt(df_filtered)['value'].value_counts() # using more than one column as an identifire 
    
df = pd.read_csv("/Users/Desktop/test.csv")  # read the csv file
df = df.transpose()
df = pd.melt(df[-2:])['value'].value_counts() #counting phyla

df.to_csv (r'/Users/Desktop/eukaryotes.csv') # the output file

我正在寻找的输出是让最后列出的组位于顶部，其计数如下

group              count
Aconoidasida       1
Apicomplexa        2
Alveolata          2
Sar                2
Viridiplantae      1
Eukaryota          3
cellular organisms 3
root               3

Answer 1

我假设每一行都包含相同的类别（例如订单、家庭、物种等）：

import pandas as pd
import numpy as np

data = {
'Column1': ['root', 'cellular organisms', 'Eukaryota', 'Sar', 'Alveolata', 'Apicomplexa', 'Aconoidasida'],
'Column2': ['root', 'cellular organisms', 'Eukaryota', 'Sar', 'Alveolata', 'Apicomplexa', ''],
'Column3': ['root', 'cellular organisms', 'Eukaryota', 'Viridiplantae', '', '', '']
}
df = pd.DataFrame(data).replace('', np.nan)

给予

              Column1             Column2             Column3
0                root                root                root
1  cellular organisms  cellular organisms  cellular organisms
2           Eukaryota           Eukaryota           Eukaryota
3                 Sar                 Sar       Viridiplantae
4           Alveolata           Alveolata                 NaN
5         Apicomplexa         Apicomplexa                 NaN
6        Aconoidasida                 NaN                 NaN

您可以遍历行并将它们的值计数计算为字典，然后合并它们（注意使用 .loc[::-1] 的反转）：

counts = df.apply(lambda row: row.value_counts().to_dict(), axis=1)
merged = {group: count for d in counts.loc[::-1] for group, count in d.items()}

给予

{'Aconoidasida': 1,
 'Alveolata': 2,
 'Apicomplexa': 2,
 'Eukaryota': 3,
 'Sar': 2,
 'Viridiplantae': 1,
 'cellular organisms': 3,
 'root': 3}

如果需要，您可以将其转换为 DataFrame：

pd.DataFrame.from_dict(merged, orient='index', columns=['count'])

给予

                    count
Aconoidasida            1
Apicomplexa             2
Alveolata               2
Sar                     2
Viridiplantae           1
Eukaryota               3
cellular organisms      3
root                    3

从头到尾计算 CSV 行中的沿袭

Counting lineage in rows in a CSV from the end to beginning

python

bioinformatics

biopython

pandas