NumPy & Pandas 자주 쓰는 문법 요약

빅데이터분석기사 실기 시험을 준비하면서, NumPy와 Pandas에서 자주 쓰는 문법들을 요약해 외우고자 cheatsheet처럼 정리해보았습니다. 문법과 그에 따른 결과를 최대한 한 줄로 보기 쉽게 만들었습니다.

NumPy

import numpy as np

a1 = np.array([[1, 2, 3], [4, 5, 6]])
print(a1)  # [[1 2 3] [4 5 6]]
print(a1.shape)  # (2, 3)
print(a1.ndim)  # 2

## Array
a2 = np.arange(1, 20, 3)  # [1 4 7 10 13 16 19]
a3 = np.zeros((2, 3))  # [[0. 0. 0.] [0. 0. 0.]]
a4 = np.ones((3, 2))  # [[1. 1.] [1. 1.] [1. 1.]]
a5 = np.zeros_like(a4)  # [[0. 0.] [0. 0.] [0. 0.]]
a6 = np.full((2, 3), 9)  # [[9 9 9] [9 9 9]]
a7 = np.random.random((3, 2))  # [[0.8 0.2] [0.7 0.5] [0.7 0.1]]
a8 = np.eye(4)  # [[1. 0. 0.] [0. 1. 0.] [0. 0. 1.]]

## Shape
a9 = np.arange(6).reshape(2, 3)  # [[0 1 2] [3 4 5]]
a10 = np.arange(6).reshape(-1, 2)  # [[0 1] [2 3] [4 5]]
a11 = np.dot(a9, a10)  # [[10 13] [28 40]]
a12 = a11.transpose()  # [[10 28] [13 40]]
a13 = a11.flatten()  # [10 13 28 40]

## Stats
print(len(a13), np.mean(a13), np.var(a13), np.std(a13))
# 4 22.75 145.688 12.07010770457331
print(np.max(a13), np.min(a13), np.median(a13), np.percentile(a13, 75))
# 40 10 20.5 31.0

Pandas

import pandas as pd

s = pd.Series([1, 3, 5, 6])
print(s.index)  # RangeIndex(start=0, stop=4, step=1)
print(s.values)  # [1 3 5 6]

## DataFrame
df1 = pd.DataFrame(
    [[1, 2], [3, 4], [5, 6]], index=["r1", "r2", "r3"], columns=["c1", "c2"]
)
df2 = pd.DataFrame({"c1": [1, 3, 5], "c2": [2, 4, 6]}, index=["r1", "r2", "r3"])
print(df1)
print(df2)
"""
    c1  c2        c1  c2
r1   1   2    r1   1   2
r2   3   4    r2   3   4
r3   5   6    r3   5   6
"""

## CSV
df3 = pd.read_csv("file_name.csv", header=None)
df3.to_csv("file_name.tsv", index=False, sep="\t")

## Summary
print(df1.shape)  # (3, 2)
print(df1.head(2))  # head rows
print(df1.info())  # columns, non-null counts, data types, memory usages
print(df1.describe())  # count, mean, std, min, max, percentiles
print(df1["c2"].value_counts())  # {2: 1, 4: 1, 6: 1}

## Stats
print(df1.sum())  # {"c1": 9, "c2": 12}
print(df1.count())  # {"c1": 3, "c2": 3}
print(df1.min())  # {"c1": 1, "c2": 2}
print(df1.max())  # {"c1": 5, "c2": 6}
print(df1.mean())  # {"c1": 3.0, "c2": 4.0}
print(df1.var())  # {"c1": 4.0, "c2": 4.0}
print(df1.std())  # {"c1": 2.0, "c2": 2.0}
print(df1.quantile(0.25))  # {"c1": 2.0, "c2": 3.0}
print(df1.corr())
"""
     c1   c2
c1  1.0  1.0
c2  1.0  1.0
"""

## GroupBy
df4 = pd.DataFrame(
    {"c1": ["a", "a", "b"], "c2": [1, 2, 3], "c3": [4, 5, 6]},
    index=["r1", "r2", "r3"],
)
print(df4.groupby("c1").mean())
"""
     c2   c3
c1
a   1.5  4.5
b   3.0  6.0
"""

## Drop
print(df4.drop(["r1", "r3"]))
print(df4.drop(["c1", "c3"], axis=1))
"""
   c1  c2  c3
r2  a   2   5
    c2
r1   1
r2   2
r3   3
"""

## Indexing
print(df1.iloc[1, 1])  # 4
print(df1.loc["r2", "c2"])  # 4
print(df1[df1["c2"] == 4])
print(df1[(df1["c2"] > 2) & (df1["c1"] < 4)])
print(df1.sort_values(by="c2", ascending=False))
"""
    c1  c2        c1  c2
r2   3   4    r2   3   4
    c1  c2
r3   5   6
r2   3   4
r1   1   2
"""

## NaN
df1["c3"] = np.nan
print(df1.isna())
print(df1.fillna(df1["c2"].mean()))
"""
       c1     c2    c3
r1  False  False  True
r2  False  False  True
r3  False  False  True
   c1  c2   c4
r1  1   2  4.0
r2  3   4  4.0
r3  5   6  4.0
"""

## Concat
print(pd.concat([df1, df2]))
print(pd.concat([df1, df2], axis=1))
"""
    c1  c2
r1   1   2
r2   3   4
r3   5   6
r1   1   2
r2   3   4
r3   5   6
    c1  c2  c1  c2
r1   1   2   1   2
r2   3   4   3   4
r3   5   6   5   6
"""