Pandas Data Objects: Series, DataFrame Creation, Indexing, CRUD Operations, and Common Functions

This article provides a comprehensive guide to pandas, a Python library for data manipulation, focusing on its two primary data structures: Series and DataFrame .

Series is a one‑dimensional labeled array. A simple creation example is:

# 通过传入一个序列给pd.Series初始化一个Series对象, 比如list

s1 = pd.Series(list("1234"))

print(s1)

0    1

1    2

2    3

3    4

dtype: object

DataFrame is a two‑dimensional table‑like structure. It can be created from a NumPy array or a dictionary:

# 通过传入一个numpy的二维数组或者dict对象给pd.DataFrame初始化一个DataFrame对象

import numpy as np

df1 = pd.DataFrame(np.random.randn(6,4))

print(df1)

# 通过dict字典

df2 = pd.DataFrame({
    'A' : 1.,
    'B' : pd.Timestamp('20130102'),
    'C' : pd.Series(1, index=list(range(4)), dtype='float32'),
    'D' : np.array([3] * 4, dtype='int32'),
    'E' : pd.Categorical(["test","train","test","train"]),
    'F' : 'foo'
})

print(df2)

Both objects have an index; for Series the index labels each element, while for DataFrame it labels rows. The index can be inspected with print(s1.index) or print(df.index) .

Selection and Query – you can view columns, rows, or subsets using .head() , .tail() , slicing, .loc (label‑based) and .iloc (position‑based). Examples:

# 查看前5行

df.head()

# 查看第10行的date和code列

df.loc[10, ["date", "code"]]

# 按位置查看第1行

df.iloc[0]

Conditional filtering uses boolean masks:

# 查看open列大于10的前5行

df[df.open > 10].head()

# 同时满足两个条件

df[(df.open > 10) & (df.open < 10.6)].head()

Modification – change index, rename columns, or apply functions:

# 将df的索引修改为date列的数据并转换为datetime

df.index = pd.to_datetime(df.date)

# 重命名列

df.columns = ["Date","Open","Close","High","Low","Volume","Code"]

# 给Open列每个数值加1

df.Open = df.Open.apply(lambda x: x+1)

Deletion – use .drop() with inplace=True if you want to modify the original object:

# 删除Open列

df.drop("Open", axis=1, inplace=True)

Common parameters and functions – control display format, obtain descriptive statistics, and handle missing values:

# 设置浮点数显示格式

pd.options.display.float_format = '{:.4f}'.format

# 统计描述

df.describe()

# 删除含NaN的行

df.dropna(how='any')

# 用固定值填充NaN

df.fillna(value=5)

Sorting and Merging – sort by index or values, concatenate rows/columns, or merge datasets:

# 按列排序

df.sort_index(axis=1).head()

# 按Open列升序排序

df.sort_values(by="Open").head()

# 行方向合并

pd.concat([df.iloc[0:2,:], df.iloc[2:4,:], df.iloc[4:9]])

Copying – because DataFrames are mutable references, use .copy() to create an independent copy before making changes.

Plotting – pandas integrates with matplotlib. Simple line and area plots can be generated with:

df[["Open","Low","High","Close"]].plot()

df[["Open","Low","High","Close"]].plot(kind="area")

Data I/O – read and write common file formats such as CSV, Excel, JSON, or even the system clipboard:

# 保存为CSV

df.to_csv("stock.csv")

# 从CSV读取并指定第一列为索引

df2 = pd.read_csv("stock.csv", index_col=0)

Practical Example: Web Log Analysis – the article demonstrates parsing an Apache access log into a DataFrame, converting the timestamp, casting numeric fields, and performing analyses such as status‑code counts, top IP addresses, and time‑series visualisations.

# 正则表达式解析日志

REGEX = HOST+SPACE+IDENTITY+SPACE+USER+SPACE+TIME+SPACE+REQUEST+SPACE+STATUS+SPACE+SIZE+SPACE+IDENTITY+USER_AGENT+SPACE

log_df = pd.DataFrame(field_lis, columns=["Host","Time","Method","Path","Protocol","status","size","User_Agent"])

# 将Time列转为datetime并设为索引

log_df.Time = log_df.Time.apply(lambda x: x.replace(":", " ", 1))
log_df.Time = pd.to_datetime(log_df.Time)
log_df.set_index('Time', inplace=True)

# 统计不同状态码出现次数

log_df.Status.value_counts()

# 绘制状态码饼图

log_df.Status.value_counts().plot(kind="pie", figsize=(10,8))

Overall, the guide equips readers with the essential pandas operations needed for data cleaning, transformation, analysis, and visualization, making it a solid reference for anyone working with tabular data in Python.