R, Python 분석과 프로그래밍의 친구 (by R Friend)

import pandas as pd

import numpy as np

# generate time series index

range = pd.date_range('2019-12-19', '2019-12-20', freq='2min')

df = pd.DataFrame(index = range)[:20]

# add 'price' columm using random number

np.random.seed(seed=1004) # for reproducibility

df['price'] = np.random.randint(low=10, high=100, size=20)

# add 'amount' column unsing random number

df['amount'] = np.random.randint(low=1, high=5, size=20)

print('Shape of df DataFrame:', df.shape)

[Out]:Shape of df DataFrame: (20, 2)

df

[Out]:

  (1) 10분 단위 구간별로 각 칼럼의 첫번째 값(first value), 마지막 값(last value) 구하기

      (select the first and last value by 10 minutes time span using pandas resample method)

# Resampling by a given time span (group)

# : first, last

df_summary = pd.DataFrame()

df_summary['price_10m_first'] = df.price.resample('10T').first()

df_summary['price_10m_last'] = df.price.resample('10T').last()

df_summary['amount_10m_first'] = df.amount.resample('10T').first()

df_summary['amount_10m_last'] = df.amount.resample('10T').last()

df_summary

  (2) 10분 단위 구간별로 숫자형 데이터의 합계, 누적 합계 구하기

     (sum, cumulative sum by 10 minutes time span using pandas resample method)

# Resampling by a given time span (group)

# sum, cumulative sum

df_summary = pd.DataFrame()

df_summary['price_10m_sum'] = df.price.resample('10T').sum()

df_summary['price_10m_cumsum'] = df.price.resample('10T').sum().cumsum()

df_summary['amount_10m_sum'] = df.amount.resample('10T').sum()

df_summary['amount_10m_cumsum'] = df.amount.resample('10T').sum().cumsum()

df_summary

  (3) 10분 단위 구간별로 최소값, 최대값, 평균, 중앙값, 범위 구하기

   (summary statistics by 10 minutes time span using pandas resample method)

# Resampling by a given time span (group)

# min, max, mean, median, range

df_summary = pd.DataFrame()

df_summary['price_10m_min'] = df.price.resample('10T').min()

df_summary['price_10m_max'] = df.price.resample('10T').max()

df_summary['price_10m_mean'] = df.price.resample('10T').mean()

df_summary['price_10m_median'] = df.price.resample('10T').median()

df_summary['price_10m_range'] = \

    df.price.resample('10T').max() - df.price.resample('10T').min()

df_summary

  (4) 10분 단위 구간별로 분산(variance), 표준편차(standard deviation) 구하기

    (variance, standard deviation by 10 minutes time span using pandas resample(() method)

# Resampling by a given time span (group)

# variance, standard deviation

df_summary = pd.DataFrame()

# sample variance 1/(N-1)*sigma(X-X_bar)^2

df_summary['price_10m_var'] = df.price.resample('10T').var() 

# sample standard deviation using sqrt(var) formula

df_summary['price_10m_stddev'] = np.sqrt(df.price.resample('10T').var())

  (5) 특정 시간 단위 구간별로 요약 통계량 구하는 사용자 정의 함수

     (User Defined Function for aggregating summary statistics by specific time span)

# UDF of Resampling by column name, time span and summary functions 

def resample_summary(ts_data, col_nm, time_span, func_list):

    import numpy as np

    import pandas as pd

    df_summary = pd.DataFrame() # blank DataFrame to store results

    # resampler with column name by time span (group by)

    resampler = ts_data[col_nm].resample(time_span)

    # aggregation functions with suffix name

    if 'first' in func_list:

        df_summary[col_nm + '_' + time_span + '_first'] = resampler.first()

    if 'last' in func_list:

        df_summary[col_nm + '_' + time_span + '_last'] = resampler.last()

    if 'sum' in func_list:

        df_summary[col_nm + '_' + time_span + '_sum'] = resampler.sum()

    if 'cumsum' in func_list:

        df_summary[col_nm + '_' + time_span + '_cumsum'] = resampler.sum().cumsum()

    if 'min' in func_list:

        df_summary[col_nm + '_' + time_span + '_min'] = resampler.min()

    if 'max' in func_list:

        df_summary[col_nm + '_' + time_span + '_max'] = resampler.max()

    if 'mean' in func_list:

        df_summary[col_nm + '_' + time_span + '_mean'] = resampler.mean()

    if 'median' in func_list:

        df_summary[col_nm + '_' + time_span + '_median'] = resampler.median()

    if 'range' in func_list:

        df_summary[col_nm + '_' + time_span + '_range'] = resampler.max() - resampler.min()

    if 'var' in func_list:

        df_summary[col_nm + '_' + time_span + '_var'] = resampler.var() # sample variance

    if 'stddev' in func_list:

        df_summary[col_nm + '_' + time_span + '_stddev'] = np.sqrt(resampler.var())

    return df_summary

func_list = ['first', 'last', 'sum', 'cumsum', 'min', 'max']

resample_summary(df, 'price', '10T', func_list)

func_list = ['mean', 'median', 'range', 'var', 'stddev']

resample_summary(df, 'price', '20T', func_list)

func_list = ['mean', 'median', 'range', 'var', 'stddev']

resample_summary(df, 'amount', '20T', func_list) # with 'amount' column

func_list = ['mean', 'median', 'range', 'var', 'stddev']

resample_summary(df, 'price', '10T', func_list)

# generate time series index

range = pd.date_range('2019-12-19', '2020-01-18', freq='2min') # one month period

df_1m = pd.DataFrame(index = range)

# add 'price' columm using random number

np.random.seed(seed=1004) # for reproducibility

df_1m['price'] = np.random.randint(low=10, high=100, size=len(df))

# add 'amount' column unsing random number

df_1m['amount'] = np.random.randint(low=1, high=5, size=len(df))

print('Shape of df_1m DataFrame:', df_1m.shape)

Shape of df_1m DataFrame: (21601, 2)

# by 1 Hour

func_list = ['first', 'sum', 'mean', 'stddev']
resample_summary(df_1m, 'price', '1H', func_list).head() # by 1 Hour

price_1H_first
price_1H_sum
price_1H_mean
price_1H_stddev
2019-12-19 00:00:00
12
1684
56.133333
25.143359
2019-12-19 01:00:00
44
1534
51.133333
24.764732
2019-12-19 02:00:00
70
1435
47.833333
25.223256
2019-12-19 03:00:00
22
1867
62.233333
24.842515
2019-12-19 04:00:00
80
1766
58.866667
23.292345

# by 1 Day

func_list = ['first', 'sum', 'mean', 'stddev']
resample_summary(df_1m, 'price', '1D', func_list).head() # by 1 Day

price_1D_first
price_1D_sum
price_1D_mean
price_1D_stddev
2019-12-19
12
39746
55.202778
25.946355
2019-12-20
26
40171
55.793056
25.547419
2019-12-21
87
39737
55.190278
26.238314
2019-12-22
65
39350
54.652778
25.675714
2019-12-23
69
39835
55.326389
26.230239

# by 1 Week

func_list = ['first', 'sum', 'mean', 'stddev']
resample_summary(df_1m, 'price', '1W', func_list) # by 1 Week

price_1W_first
price_1W_sum
price_1W_mean
price_1W_stddev
2019-12-22
12
159004
55.209722
25.842990
2019-12-29
69
272943
54.155357
26.084089
2020-01-05
72
274740
54.511905
25.840425
2020-01-12
41
276563
54.873611
26.295806
2020-01-19
55
197090
54.732019
25.984207

# by 2 Weeks

func_list = ['first', 'sum', 'mean', 'stddev']
resample_summary(df_1m, 'price', '2W', func_list) # by 2 Week

price_2W_first
price_2W_sum
price_2W_mean
price_2W_stddev
2019-12-22
12
159004
55.209722
25.842990
2020-01-05
69
547683
54.333631
25.961867
2020-01-19
41
473653
54.814605
26.164988

# by 1 Month

func_list = ['first', 'sum', 'mean', 'stddev']
resample_summary(df_1m, 'price', '1M', func_list) # by 1 Month

price_1M_first
price_1M_sum
price_1M_mean
price_1M_stddev
2019-12-31
12
510036
54.491026
25.912189
2020-01-31
48
670304
54.758925
26.117109

  (6) 10분 단위 구간별 수량 가중 평균 가격 구하기 

   (amount-weighted average of price by 10 minutes time span using pandas resample method)

# function: weighted average

# 각 시간대의 수량가중평균가격(sum(price*amount)/sum(amount))

# (*가중평균은 특정 시간대에 발생한 모든 구입건의 구입가격*구입수량 합/전체 구입수량 합)

df_summary = pd.DataFrame()

df['price_mult_amt'] = df['price']*df['amount']

df_summary['price_10m_amount_weighted_avg'] = \

    df.price_mult_amt.resample('10T').sum() / df.amount.resample('10T').sum()

df_summary

  (7) 10분 단위 구간별 집계/요약 통계량 결과를 csv 파일로 내보내기 

   (exporting summary results by 10 minutes time span into 'csv file' using pandas to_csv() method)

# summary statistics using resample_summary() User Defiened Function, refer to (5)

func_list = ['first', 'last', 'sum', 'cumsum', 'min', 'max', 

            'mean', 'median', 'range', 'var', 'stddev']

df_summary = resample_summary(df, 'price', '10T', func_list)

# amount-weighted average of price, refer to (6)

df['price_mult_amt'] = df['price']*df['amount']

df_summary['price_10m_amount_weighted_avg'] = \

    df.price_mult_amt.resample('10T').sum()/ df.amount.resample('10T').sum()

# export df_summary DataFrame into csv file

import os

work_dir = os.getcwd() # current working directory

file_path = os.path.join(work_dir, 'df_summary.csv')

df_summary.to_csv(file_path

                  , index=True             # include index

                  , na_rep='NaN'           # representation of missing value

                  , float_format = '%.2f') # 2 decimal places

  1. PostgreSQL로 그룹별로 특정 칼럼의 행을 하나씩 내리기, 올리기 

    (lag, lead a row by group using PostgreSQL lag(), lead() window function)

DROP TABLE IF EXISTS ts;

CREATE TABLE ts (

    dt date not null

    , id text not null  

    , val numeric not null

);

INSERT INTO ts VALUES 

  ('2019-12-01', 'a', 5)

, ('2019-12-02', 'a', 6)

, ('2019-12-03', 'a', 7)

, ('2019-12-04', 'a', 8)

, ('2019-12-01', 'b', 13)

, ('2019-12-02', 'b', 14)

, ('2019-12-03', 'b', 15)

, ('2019-12-04', 'b', 16);

SELECT * FROM ts ORDER BY id, dt;

-- lead() windows function

SELECT 

    *

    , LAG(val, 1, NULL) OVER (PARTITION BY id ORDER BY dt) AS val_lag_1

    , LEAD(val, 1, NULL) OVER (PARTITION BY id ORDER BY dt) AS val_lead_2

FROM ts;

DROP TABLE IF EXISTS ts_lag_lead;

CREATE TABLE ts_lag_lead AS (

SELECT 

    *

    , LAG(val, 1, NULL) OVER (PARTITION BY id ORDER BY dt) AS val_lag_1

    , LEAD(val, 1, NULL) OVER (PARTITION BY id ORDER BY dt) AS val_lead_2

FROM ts

);

SELECT * FROM ts_lag_lead ORDER BY id, dt;

  2. Python pandas 로 DataFrame 내 그룹별 특정 칼럼의 행을 하나씩 내리기, 올리기 

     (shift a row by group using Python pandas library)

import pandas as pd

ts = pd.DataFrame({'dt': ['2019-12-01', '2019-12-02', '2019-12-03', '2019-12-04', 

                          '2019-12-01', '2019-12-02', '2019-12-03', '2019-12-04'], 

                  'id': ['a', 'a', 'a', 'a', 'b', 'b', 'b', 'b'], 

                  'val': [5, 6, 7, 8, 13, 14, 15, 16]})

ts

# lag a row by group 'id'

ts['val_lag_1'] =  ts.sort_values(by='dt', ascending=True).groupby('id')['val'].shift(1)

# lead a row by group 'id'

ts['val_lead_1'] = ts.sort_values(by='dt', ascending=False).groupby('id')['val'].shift(1)

ts.sort_values(by=['id', 'dt'])

  3. R dplyr 로 dataframe 내 그룹별 특정 칼럼의 행을 하나씩 내리기, 올리기 

     (lag, lead a row by group using R dplyr library)

#install.packages("dplyr")

library(dplyr)

dt <- c(rep(c('2019-12-01', '2019-12-02', '2019-12-03', '2019-12-04'), 2))

id <- c(rep('a', 4), rep('b', 4)) 

val <- c(5, 6, 7, 8, 13, 14, 15, 16)

ts <- data.frame(dt, id, val)

ts

ts <- ts %>% 

    arrange(dt) %>%

    group_by(id)  %>% 

    mutate(val_lag_1 = lag(val, 1), 

          val_lead_1 = lead(val, 1))

arrange(ts, id, dt)

 (1) Jupyter Notebook cell 너비 설정하기 (setting cell width in Jupyter Notebook)

from IPython.core.display import display, HTML

display(HTML("<style>.container { width: 50% !important; }</style>"))

 (2) Jupyter Notebook에서 DataFrame 칼럼 최대 너비 설정하기 

     (setting the max-width of DataFrame's column in Jupyter Notebook)

import pandas as pd

pd.set_option('display.max.colwidth', 10)

df = pd.DataFrame({'a': [100000000.0123, 20000000.54321], 

                  'b': ['abcdefghijklmnop', 'qrstuvwxyz']})

df

 (3) Jupyter Notebook에서 DataFrame 내 텍스트를 왼쪽으로 정렬하기 

     (align text of pandas DataFrame to left in Jupyter Notebook)

dfStyler = df.style.set_properties(**{'text-align': 'left'})

dfStyler.set_table_styles([dict(selector='th', 

                                props=[('text-align', 'left')])])

 (4) Jupyter Notebook에서 DataFrame 소수점 자리수 설정하기

     (setting the decimal point format of pandas DataFrame in Jupyter Notebook)

import pandas as pd

pd.options.display.float_format = '{:.2f}'.format

df

 (5) Jupyter Notebook에서 matplotlib plot 기본 옵션 설정하기 

    (setting figure size, line width, color, grid of matplotlib plot in Jupyter Notebook)

# matplotlib setting

import matplotlib.pylab as plt

%matplotlib inline

plt.rcParams["figure.figsize"] = (6, 5)

plt.rcParams["lines.linewidth"] = 2

plt.rcParams["lines.color"] = 'r'

plt.rcParams["axes.grid"] = True

# simple plot

x = [1, 2, 3, 4, 5]

y = [2, 3.5, 5, 8, 9]

plt.plot(x, y)

plt.show()

  1. Jupyter Notebook 사용할 수 있도록 IRKernel 설치하기

MacBook-Pro:~ ihongdon$

MacBook-Pro:~ ihongdon$ R

R version 3.6.0 (2019-04-26) -- "Planting of a Tree"

Copyright (C) 2019 The R Foundation for Statistical Computing

Platform: x86_64-apple-darwin15.6.0 (64-bit)

>

> install.packages('devtools')

--- 현재 세션에서 사용할 CRAN 미러를 선택해 주세요 ---

SEOUL

경고: 저장소 https://cran.seoul.go.kr/bin/macosx/el-capitan/contrib/3.6에 대한 인덱스에 접근할 수 없습니다:

  URL 'https://cran.seoul.go.kr/bin/macosx/el-capitan/contrib/3.6/PACKAGES'를 열 수 없습니다

소스형태의 패키지 ‘devtools’(들)를 설치합니다.

URL 'https://cran.seoul.go.kr/src/contrib/devtools_2.2.1.tar.gz'을 시도합니다

Content type 'application/x-gzip' length 372273 bytes (363 KB)

==================================================

downloaded 363 KB

* installing *source* package ‘devtools’ ...

** 패키지 ‘devtools’는 성공적으로 압축해제되었고, MD5 sums 이 확인되었습니다

** using staged installation

** R

** inst

** byte-compile and prepare package for lazy loading

** help

*** installing help indices

*** copying figures

** building package indices

** installing vignettes

** testing if installed package can be loaded from temporary location

** testing if installed package can be loaded from final location

** testing if installed package keeps a record of temporary installation path

* DONE (devtools)

다운로드한 소스 패키지들은 다음의 위치에 있습니다

‘/private/var/folders/6q/mtq6ftrj6_z4txn_zsxcfyxc0000gn/T/Rtmpg1bePN/downloaded_packages’

>

>

> devtools::install_github('IRkernel/IRkernel')

Downloading GitHub repo IRkernel/IRkernel@master

'/usr/local/bin/git' clone --depth 1 --no-hardlinks --recurse-submodules https://github.com/jupyter/jupyter_kernel_test.git /var/folders/6q/mtq6ftrj6_z4txn_zsxcfyxc0000gn/T//Rtmpg1bePN/remotes36921150b621/IRkernel-IRkernel-67592db/tests/testthat/jkt

'/var/folders/6q/mtq6ftrj6_z4txn_zsxcfyxc0000gn/T//Rtmpg1bePN/remotes36921150b621/IRkernel-IRkernel-67592db/tests/testthat/jkt'에 복제합니다...

remote: Enumerating objects: 12, done.

remote: Counting objects: 100% (12/12), done.

remote: Compressing objects: 100% (11/11), done.

remote: Total 12 (delta 1), reused 3 (delta 0), pack-reused 0

오브젝트 묶음 푸는 중: 100% (12/12), 완료.

'/usr/local/bin/git' clone --depth 1 --no-hardlinks --recurse-submodules https://github.com/flying-sheep/ndjson-testrunner.git /var/folders/6q/mtq6ftrj6_z4txn_zsxcfyxc0000gn/T//Rtmpg1bePN/remotes36921150b621/IRkernel-IRkernel-67592db/tests/testthat/njr

'/var/folders/6q/mtq6ftrj6_z4txn_zsxcfyxc0000gn/T//Rtmpg1bePN/remotes36921150b621/IRkernel-IRkernel-67592db/tests/testthat/njr'에 복제합니다...

remote: Enumerating objects: 10, done.

remote: Counting objects: 100% (10/10), done.

remote: Compressing objects: 100% (8/8), done.

remote: Total 10 (delta 0), reused 6 (delta 0), pack-reused 0

오브젝트 묶음 푸는 중: 100% (10/10), 완료.

Skipping 1 packages ahead of CRAN: htmltools

   checking for file ‘/private/var/folders/6q/mtq6ftrj6_z4txn_zsxcfyxc0000gn/T/Rtmpg1bePN/remotes36921150b621/IRkernel✔  checking for file ‘/private/var/folders/6q/mtq6ftrj6_z4txn_zsxcfyxc0000gn/T/Rtmpg1bePN/remotes36921150b621/IRkernel-IRkernel-67592db/DESCRIPTION’

─  preparing ‘IRkernel’:

✔  checking DESCRIPTION meta-information ...

─  checking for LF line-endings in source and make files and shell scripts

─  checking for empty or unneeded directories

   Removed empty directory ‘IRkernel/example-notebooks’

─  building ‘IRkernel_1.0.2.9000.tar.gz’

* installing *source* package ‘IRkernel’ ...

** using staged installation

** R

** inst

** byte-compile and prepare package for lazy loading

** help

*** installing help indices

** building package indices

** testing if installed package can be loaded from temporary location

** testing if installed package can be loaded from final location

** testing if installed package keeps a record of temporary installation path

* DONE (IRkernel)

>

>

> IRkernel::installspec()

[InstallKernelSpec] Installed kernelspec ir in /Users/ihongdon/Library/Jupyter/kernels/ir

>

 Anaconda\Lib\site-packages\jupyter_client

 C:\Users\Anaconda3\Scripts 

  2. Jupyter Notebook에서 R 사용해보기

>

>

[1]+  Stopped                 R

MacBook-Pro:~ ihongdon$

MacBook-Pro:~ ihongdon$ conda info -e

# conda environments:

#

base                  *  /Users/ihongdon/anaconda3

py2.7_tf1.4              /Users/ihongdon/anaconda3/envs/py2.7_tf1.4

py3.5_tf1.4              /Users/ihongdon/anaconda3/envs/py3.5_tf1.4

py3.6_tf2.0              /Users/ihongdon/anaconda3/envs/py3.6_tf2.0

MacBook-Pro:~ ihongdon$

MacBook-Pro:~ ihongdon$

MacBook-Pro:~ ihongdon$ source activate py3.6_tf2.0

(py3.6_tf2.0) MacBook-Pro:~ ihongdon$

(py3.6_tf2.0) MacBook-Pro:~ ihongdon$ jupyter notebook

<CD>

</CATALOG> 

- source: https://www.w3schools.com/xml/cd_catalog.xml -

import pandas as pd

import xml.etree.ElementTree as ET

import sys

if sys.version_info[0] == 3:

    from urllib.request import urlopen

else:

    from urllib import urlopen

# If you are using Python 3.x version, then ImportError will be raised as below

from urllib import urlopen

---------------------------------------------------------------------------
ImportError                               Traceback (most recent call last)
<ipython-input-2-dbf1dbb53f94> in <module>()
----> 1 from urllib import urlopen

ImportError: cannot import name 'urlopen'

url = "https://www.w3schools.com/xml/cd_catalog.xml"

response = urlopen(url).read()

xtree = ET.fromstring(response)

xtree

<Element 'CATALOG' at 0x00000219E77DCCC8>

rows = []

# iterate through each node of the tree

for node in xtree:

    n_title = node.find("TITLE").text

    n_artist = node.find("ARTIST").text

    n_country = node.find("COUNTRY").text

    n_company = node.find("COMPANY").text

    n_price = node.find("PRICE").text

    n_year = node.find("YEAR").text

    rows.append({"title": n_title, 

                 "artist": n_artist, 

                 "country": n_country, 

                 "company": n_company, 

                 "price": n_price, 

                 "year": n_year})

# convert XML data to pandas DataFrame

columns = ["title", "artist", "country", "company", "price", "year"]

catalog_cd_df = pd.DataFrame(rows, columns = columns)

catalog_cd_df.head(10)

catalog_cd_df.dtypes

title      object
artist     object
country    object
company    object
price      object
year       object
dtype: object

import numpy as np

catalog_cd_df = catalog_cd_df.astype({'price': np.float, 

                                      'year': int})

catalog_cd_df.dtypes

title       object
artist      object
country     object
company     object
price      float64
year         int32
dtype: object

country_mean = catalog_cd_df.groupby('country').price.mean()

country_mean

country
EU        9.320000
Norway    7.900000
UK        8.984615
USA       9.385714
Name: price, dtype: float64

country_mean_df = pd.DataFrame(country_mean).reset_index()

import seaborn as sns

sns.barplot(x='country', y='price', data=country_mean_df)

plt.show()

 (1) API 웹 사이트에서 JSON 포맷 자료를 Python으로 읽어오기

# parse a JSON string using json.loads() method : returns a dictionary

import json

import urllib

import pandas as pd

# API request to the URL

import sys

if sys.version_info[0] == 3:

    from urllib.request import urlopen # for Python 3.x

else:

    from urllib import urlopen           # for Python 2.x

with urlopen("http://api.nobelprize.org/v1/prize.json") as url:

    novel_prize_json_file = url.read()

# ImportError: cannot import name 'urlopen' at python 3.x

from urllib import urlopen # It's only for Python 2.x. It's not working at Python 3.x

with urlopen("http://api.nobelprize.org/v1/prize.json") as url:

    novel_prize_json_file = url.read()

---------------------------------------------------------------------------
ImportError                               Traceback (most recent call last)
<ipython-input-1-81c8fae1a1fd> in <module>()
      1 # API request to the URL
----> 2 from urllib import urlopen
      3 
      4 with urlopen("http://api.nobelprize.org/v1/prize.json") as url:
      5     novel_prize_json_file = url.read()

ImportError: cannot import name 'urlopen'

# decoding to python object

novel_prize_json = json.loads(novel_prize_json_file.decode('utf-8'))

# decoding TypeError. decode using decode('utf-8')

novel_prize_json = json.loads(novel_prize_json_file.decode)

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-9-1fda68be6386> in <module>()
      1 # decoding TypeError
----> 2 novel_prize_json = json.loads(novel_prize_json_file.decode)

C:\Users\admin\Anaconda3\lib\json\__init__.py in loads(s, encoding, cls, object_hook, parse_float, parse_int, parse_constant, object_pairs_hook, **kw)
    310     if not isinstance(s, str):
    311         raise TypeError('the JSON object must be str, not {!r}'.format(
--> 312                             s.__class__.__name__))
    313     if s.startswith(u'\ufeff'):
    314         raise JSONDecodeError("Unexpected UTF-8 BOM (decode using utf-8-sig)",

TypeError: the JSON object must be str, not 'builtin_function_or_method'

novel_prize_json.keys()

dict_keys(['prizes'])

novel_prize_json['prizes'][0].keys()

dict_keys(['laureates', 'year', 'overallMotivation', 'category'])

novel_prize_json['prizes'][0]

{'category': 'physics',
 'laureates': [{'firstname': 'Arthur',
   'id': '960',
   'motivation': '"for the optical tweezers and their application to biological systems"',
   'share': '2',
   'surname': 'Ashkin'},
  {'firstname': 'Gérard',
   'id': '961',
   'motivation': '"for their method of generating high-intensity, ultra-short optical pulses"',
   'share': '4',
   'surname': 'Mourou'},
  {'firstname': 'Donna',
   'id': '962',
   'motivation': '"for their method of generating high-intensity, ultra-short optical pulses"',
   'share': '4',
   'surname': 'Strickland'}],
 'overallMotivation': '"for groundbreaking inventions in the field of laser physics"', 

'year': '2018'} 

print(json.dumps(novel_prize_json['prizes'][0], indent=4))

{
    "laureates": [
        {
            "share": "2",
            "id": "960",
            "surname": "Ashkin",
            "motivation": "\"for the optical tweezers and their application to biological systems\"",
            "firstname": "Arthur"
        },
        {
            "share": "4",
            "id": "961",
            "surname": "Mourou",
            "motivation": "\"for their method of generating high-intensity, ultra-short optical pulses\"",
            "firstname": "G\u00e9rard"
        },
        {
            "share": "4",
            "id": "962",
            "surname": "Strickland",
            "motivation": "\"for their method of generating high-intensity, ultra-short optical pulses\"",
            "firstname": "Donna"
        }
    ],
    "year": "2018",
    "overallMotivation": "\"for groundbreaking inventions in the field of laser physics\"",
    "category": "physics"
}

print(json.dumps(novel_prize_json['prizes'][0], indent=4, sort_keys=True))

{
    "category": "physics",
    "laureates": [
        {
            "firstname": "Arthur",
            "id": "960",
            "motivation": "\"for the optical tweezers and their application to biological systems\"",
            "share": "2",
            "surname": "Ashkin"
        },
        {
            "firstname": "G\u00e9rard",
            "id": "961",
            "motivation": "\"for their method of generating high-intensity, ultra-short optical pulses\"",
            "share": "4",
            "surname": "Mourou"
        },
        {
            "firstname": "Donna",
            "id": "962",
            "motivation": "\"for their method of generating high-intensity, ultra-short optical pulses\"",
            "share": "4",
            "surname": "Strickland"
        }
    ],
    "overallMotivation": "\"for groundbreaking inventions in the field of laser physics\"",
    "year": "2018"
}

 (2) JSON 포맷 데이터를 pandas DataFrame으로 만들기

novel_prize_physics = pd.DataFrame(novel_prize_json['prizes'][0]["laureates"])

novel_prize_physics

novel_prize_physics = pd.DataFrame(novel_prize_json['prizes'][0]["laureates"], 

                                   columns = ['id', 'firstname', 'surname', 'share', 'motivation'])

novel_prize_physics

# make a dictionary

pgm_lang = {

    "java": 20, 

    "javascript": 8, 

    "c": 7,  

    "r": 4, 

    "python": 28 } 

 (1) 키를 기준으로 오름차순 정렬 (sort by key in ascending order): sorted()

sorted(pgm_lang.keys())

['c', 'java', 'javascript', 'python', 'r']

sorted(pgm_lang.items())

[('c', 7), ('java', 20), ('javascript', 8), ('python', 28), ('r', 4)]

for key, value in sorted(pgm_lang.items()):

    print(key, ":", value)

c : 7
java : 20
javascript : 8
python : 28
r : 4

# list comprehension

[print(key, ":", value) for (key, value) in sorted(pgm_lang.items())]

c : 7
java : 20
javascript : 8
python : 28
r : 4

[None, None, None, None, None]

# sort a dictionary by custom key function (eg. by the length of key strings)

pgm_lang_len = sorted(pgm_lang.items(), key = lambda item: len(item[0])) # key: [0]

for key, value in pgm_lang_len:

    print(key, ":", value)

c : 7
r : 4
java : 20
python : 28
javascript : 8

 (2) 키를 기준으로 내림차순 정렬 (sort by key in descending order): reverse=True

# sorting in reverse order

sorted(pgm_lang.keys(), reverse=True)