python - 使用固定宽度的列保存数据框

Question

我需要编辑使用固定宽度列的程序的输入。我设法正确加载它并找出我需要的内容，但我不知道如何保存编辑后的文件以保持相同的列宽。

import pandas as pd

file = pd.read_fwf('file.inp', 
                  colspecs = [(0, 6), (6, 11), (11, 16), (16, 20), (20, 22), 
                              (22, 26), (26, 38), (38, 46), (46, 54), (54, 61),
                              (61, 68), (68, 90)])

问题在于列之间的空格数量或数据字符串的长度各不相同，因此我不能简单地在列之间添加固定数量的空格。

要编辑的文件如下所示:

ATOM    873  N   ALA A  59      41.629  23.754-163.394  1.00 12.93           N
ATOM   5089  NH1 ARG A 315      21.344 -13.371 187.612  1.00 66.09           N1+
ATOM   7839 H5''   A B   3      31.406  -4.882-165.817  1.00 16.98           H 
HETATM 7766 H161 G3A B   1      42.941   1.714-165.146  1.00 14.70           H  

它是数字、字符串、特殊字符和列粘合在一起的混合物。

Answer 1

正如我评论的那样，我在 pandas 中找不到 write_fdf 方法。但是，我认为您可以使用 tabulate 实现您想要的效果。我将以下代码基于 this post我还没有运行它:

import pandas as pd
from tabulate import tabulate


df = pd.read_fwf('file.inp', 
                  colspecs = [(0, 6), (6, 11), (11, 16), (16, 20), (20, 22), 
                              (22, 26), (26, 38), (38, 46), (46, 54), (54, 61),
                              (61, 68), (68, 90)])


with open("...", "w") as f:
    f.write(
        tabulate(
            [list(row) for row in df.values],
            tablefmt="plain"
        )
    )

再次注意，我没有运行这个，只是为了演示如何使用制表:

• 我尚未传递 headers 参数，您可以在需要时使用该参数
• 我使用“plain”格式来避免任何类似表格的装饰

有关制表的更多信息 here

更新:我意识到输出需要与特定列对齐，因此以下内容似乎足够接近(可能需要进行较小的规范编辑):

import pandas as pd
import sys


# Mock data
lst = [
    ["ATOM", 873, "N", "ALA", "A", 59, 41.629, 23.754, -163.394, 1.00, 12.93, "N"],
    ["ATOM", 5089, "NH1", "ARG", "A", 315, 21.344, -13.371, 187.612, 1.00 ,66.09, "N1+"],
    ["ATOM", 7839, "H5''", "A", "B", 3, 31.406, -4.882, -165.817, 1.00, 16.98, "H" ],
    ["HETATM", 7766, "H161", "G3A", "B", 1, 42.941, 1.714, -165.146, 1.00, 14.70, "H"],
]

# NOTE the spaces at the end, only when needed
colspecs = [
    "{: <6} ", # left, width=6
    "{: >4} ", # right, width=4
    "{: >4} ",
    "{: >3} ",
    "{: >1} ",
    "{: >3} ",
    "{: >11} ",
    "{: >7}",
    "{: >8} ",
    "{: >5} ",
    "{: <15} ",
    "{: <3}",
]


def write_fdf(fpath, pd, specs):
    """
    Write a Pandas dataframe in fixed width column format with the given
    column specs

    Args:
        fpath: File path
        ps: Dataframe
        specs: A list of python formats
    """
    with open(fpath, "w") as f:
        for _, row in df.iterrows():
            for idx, value in enumerate(row):
                sys.stdout.write(specs[idx].format(value))
                f.write(specs[idx].format(value))

            f.write("\n")
            print("")


df = pd.DataFrame(lst)
write_fdf("/tmp/out.dat", pd, colspecs)

输出:

$ python ~/tmp/test.py
ATOM    873    N ALA A  59      41.629  23.754-163.394   1.0 12.93           N
ATOM   5089  NH1 ARG A 315      21.344 -13.371 187.612   1.0 66.09           N1+
ATOM   7839 H5''   A B   3      31.406  -4.882-165.817   1.0 16.98           H
HETATM 7766 H161 G3A B   1      42.941   1.714-165.146   1.0 14.7            H


$ cat /tmp/out.dat
ATOM    873    N ALA A  59      41.629  23.754-163.394   1.0 12.93           N
ATOM   5089  NH1 ARG A 315      21.344 -13.371 187.612   1.0 66.09           N1+
ATOM   7839 H5''   A B   3      31.406  -4.882-165.817   1.0 16.98           H
HETATM 7766 H161 G3A B   1      42.941   1.714-165.146   1.0 14.7            H

但是，我发现格式中有一个异常(exception)，这似乎违反了规则:row=0，col=2:“N”似乎在您的示例中居中