我会这样做：

import csv, re

WORD, N = "aaaa", 1

pattern = (
    rf"((?:\S+ +){{0,{N}}}\S*"  # word(s) before target
    rf"\b{re.escape(WORD)}\b"   # the targeted word
    rf"\S*(?: +\S+){{0,{N}}})"  # word(s) after target
)

with open("input.csv", "r") as inpf, open("output.csv", "w") as outf:

    data = [
        {**line, "info": re.search(pattern, line["remarks"]).group()
        if re.search(pattern, line["remarks"]) else "NA"}
        for line in csv.DictReader(inpf)
    ]

    writer = csv.DictWriter(outf, fieldnames=data[0].keys())
    writer.writeheader()
    writer.writerows(data)

Regex：[ demo ]
或者使用pandas的等价物（* 使用与上面相同的 * pattern *）：

#pip install pandas
import pandas as pd

(
    pd.read_csv("input.csv")
      .assign(info=lambda x: x["remarks"]
              .str.extract(pattern, expand=False).fillna("NA"))
      .to_csv("output.csv", index=False)
)

output.csv的最后2个字段（* 以表格格式 *）：

remarks                  info
0  xxxyyyy zzzz aaaa bbbb ccc dd erfdefgrgesrg        zzzz aaaa bbbb
1  xxxyyyy zzzz aaaa bbbb ccc dd erfdefgrgesrg        zzzz aaaa bbbb
2                   aaaa bbbbb wwww dddd zzzzz            aaaa bbbbb
3  xxxyyyy zzzz aaaa bbbb ccc dd erfdefgrgesrg        zzzz aaaa bbbb
4  xxxyyyy zzzz aaaa bbbb ccc dd erfdefgrgesrg        zzzz aaaa bbbb
5            chuifheurfhr asbfswgfduwsegf aaaa  asbfswgfduwsegf aaaa
6                        loezlerl oooezp bbbll                    NA

使用input.csv：

id,f_name,1_name,fc_id,remarks
1,Raj,Sharma,1,xxxyyyy zzzz aaaa bbbb ccc dd erfdefgrgesrg
1,Raj,Sharma,2,xxxyyyy zzzz aaaa bbbb ccc dd erfdefgrgesrg
1,Raj,Sharma,2,aaaa bbbbb wwww dddd zzzzz
2,Ram,Kapoor,1,xxxyyyy zzzz aaaa bbbb ccc dd erfdefgrgesrg
2,Ram,Kapoor,2,xxxyyyy zzzz aaaa bbbb ccc dd erfdefgrgesrg
3,Raju,Verma,3,chuifheurfhr asbfswgfduwsegf aaaa
4,XXXX,Yyyyy,4,loezlerl oooezp bbbll # <-- added by me

• 更新 *：

如果你需要的目标不仅仅是一个单词，你可以尝试下面的模式：

WORDS, N = ["aaaa", "tttt"], 1

pattern = (
    rf"((?:\S+ +){{0,{N}}}\S*"
    fr"\b(?:{'|'.join(map(re.escape, WORDS))})\b"
    rf"\S*(?: +\S+){{0,{N}}})"
)