How to Parse Level‑Up Logs with Python and Pandas – Full Code Explained

Introduction

Hello, I am PiPi. A few days ago a member of the Python Ultraman group asked a Python interview question. Below is the full code version of the solution, shared for fans who prefer a direct code snippet.

Implementation Process

Previously we posted an image version of the code, but the image was blurry and typing it manually was tedious. Here is the clean, runnable code:

import pandas as
import re
import json

file = "login.txt"
df1 = pd.read_table(file, engine="python", header=None)
df1.columns = ["col1"]
df2 = pd.DataFrame(columns=["日期", "id", "lv"])

k = 0
dic1 = {}
for i in df1.index:
    if "levelup" in df1.at[i, "col1"]:
        k += 1
        df2.at[k, "日期"] = df1.at[i, "col1"][1:11]
        pattern = r"{[^}]+}"
        match = re.search(pattern, df1.at[i, "col1"])
        if match:
            dic1 = json.loads(match.group())
            df2.at[k, "id"] = dic1["id"]
            df2.at[k, "lv"] = dic1["to_lv"]

df2.drop_duplicates(subset={"日期", "id"}, keep="last", inplace=True)
print(df2)
df2_grouped = df2.groupby(["日期", "lv"]).size().reset_index(name="Count")
print(df2_grouped)

The solution works by iterating over the log file, checking each line for the keyword "levelup", extracting the timestamp, parsing the embedded JSON for the user ID and target level, and then aggregating the data. Duplicate entries are removed, and the final DataFrame is grouped by date and level.

Conclusion

This article presents a Python interview programming problem, provides a detailed analysis, and shares the complete code implementation to help readers solve the task efficiently.