Adding source code files for Chapter 6

2021-03-13 19:25:45 +04:00
parent e662762d09
commit 9589675a50
45 changed files with 625 additions and 0 deletions
@@ -0,0 +1,33 @@
+# pandastrick1.py
+import pandas as pd
+
+weekly_data = {'day':['Monday','Tuesday', 'Wednesday', 'Thursday',
+                      'Friday', 'Saturday', 'Sunday'],
+                'temp':[40, 33, 42, 31, 41, 40, 30],
+                'condition':['Sunny','Cloudy','Sunny','Rainy','Sunny',
+                        'Cloudy','Rainy']
+        }
+df = pd.DataFrame(weekly_data)
+
+#Replacing a numeric value of 40 with 39 across DF
+df.replace(40,39, inplace=True)
+
+#Replacing string Sunny with Sun across DF
+df.replace("Sunny","Sun",inplace=True)
+
+#Replacing strings starting with Cl with Cloud across DF
+df.replace(to_replace="^Cl.*",value="Cloud", inplace=True,regex=True)
+
+#Replacing Day names using a list across DF
+df.replace(["Monday","Tuesday"],["monday","tuesday"], inplace=True)
+
+#Replacing Day names using a single dict across DF
+df.replace({"Wednesday":"wednesday","Thursday":"thursday"}, inplace=True)
+
+#Replacing Day names using dict for column and value
+df.replace({"day":"Friday"}, {"day":"friday"}, inplace=True)
+
+#Replacing name of two days using dicts for column day
+df.replace({"day":{"Saturday":"saturday", "Sunday":"sunday"},
+            "condition":{"Rainy":"Rain"}}, inplace=True)
+print(df)
@@ -0,0 +1,17 @@
+# pandastrick2.py
+import pandas as pd
+
+weekly_data = {'day':['Monday','Tuesday', 'Wednesday', 'Thursday',
+                      'Friday', 'Saturday', 'Sunday'],
+                'temp':[40, 33, 42, 31, 41, 40, 30],
+                'condition':['Sunny,','_Cloudy ','Sunny','Rainy',
+                             '--Sunny.','Cloudy.','Rainy']
+        }
+df = pd.DataFrame(weekly_data)
+print(df)
+
+df["condition"] = df["condition"].map(
+                lambda x: x.lstrip('_- ').rstrip(',. '))
+
+df["temp_F"] = df["temp"].apply(lambda x: 9/5*x+32 )
+print(df)
@@ -0,0 +1,13 @@
+# pandastrick3.py
+import pandas as pd
+
+weekly_data = {'day':['Monday','Tuesday', 'Wednesday', 'Thursday',
+                      'Friday', 'Saturday', 'Sunday'],
+                'temp':[40, 33, 42, 31, 41, 40, 30],
+                'condition':['Sunny','Cloudy','Sunny','Rainy','Sunny',
+                        'Cloudy','Rainy']
+        }
+df = pd.DataFrame(weekly_data)
+
+print(df[(df.temp >= 30) & (df.temp<=40)])
+print(df[df.temp.between(30,40)])
@@ -0,0 +1,13 @@
+# pandastrick4.py
+import pandas as pd
+
+weekly_data = {'day':['Monday','Tuesday', 'Wednesday', 'Thursday',
+                      'Friday', 'Saturday', 'Sunday'],
+                'temp':[40, 33, 42, 31, 41, 40, 30],
+                'condition':['Sunny','Cloudy','Sunny','Rainy','Sunny',
+                        'Cloudy','Rainy']
+        }
+df = pd.DataFrame(weekly_data)
+
+print(df[(df.condition == 'Rainy') | (df.condition == 'Sunny')])
+print(df[df['condition'].str.contains('Rainy|Sunny')])
@@ -0,0 +1,17 @@
+# pandastrick5.py
+import pandas as pd
+import numpy as np
+pd.set_option('display.max_columns', None)
+
+weekly_data = {'day':['Monday','Tuesday', 'Wednesday', 'Thursday',
+                      'Friday', 'Saturday', 'Sunday'],
+                'temp':[40, 33, 42, 31, 41, 40, 30],
+                'condition':['Sunny','Cloudy','Sunny','Rainy','Sunny',
+                        'Cloudy','Rainy']
+        }
+df = pd.DataFrame(weekly_data)
+
+print(df.describe())
+print(df.describe(include="all"))
+print(df.describe(percentiles=np.arange(0, 1, 0.1)))
+print(df.groupby('condition').describe(percentiles=np.arange(0, 1, 0.1)))
@@ -0,0 +1,6 @@
+#between function
+
+import pandas as pd
+
+df = pd.read_csv("weekly_weather.csv")
+print(df)
@@ -0,0 +1,13 @@
+import pandas as pd
+import numpy as np
+
+df = pd.DataFrame(np.array([[1, 2, 3], [4, 5, 6]]))
+print(df)
+
+data = np.array([['', 'Col1', 'Col2'],
+                 ['Row1', 1, 2],
+                 ['Row2', 3, 4]])
+
+print(pd.DataFrame(data=data[1:, 1:],
+                   index=data[1:, 0],
+                   columns=data[0, 1:]))
@@ -0,0 +1,15 @@
+# pandas1.py
+import pandas as pd
+
+weekly_data = {'day':['Monday','Tuesday', 'Wednesday', 'Thursday',
+                      'Friday', 'Saturday', 'Sunday'],
+                'temp':[40, 33, 42, 31, 41, 40, 30],
+                'condition':['Sunny','Cloudy','Sunny','Rain','Sunny',
+                        'Cloudy','Rain']
+        }
+
+df = pd.DataFrame(weekly_data)
+print(df)
+
+df1 = df.set_index('day')
+print(df1)
@@ -0,0 +1,13 @@
+# pandas2.py
+import pandas as pd
+
+weekly_data = {'day':['Monday','Tuesday', 'Wednesday', 'Thursday',
+                      'Friday', 'Saturday', 'Sunday'],
+                'temp':[40, 33, 42, 31, 41, 40, 30],
+                'condition':['Sunny','Cloudy','Sunny','Rain','Sunny',
+                        'Cloudy','Rain']
+        }
+
+df = pd.DataFrame(weekly_data)
+df.index = ['MON', 'TUE','WED','THU','FRI','SAT','SUN']
+print(df)
@@ -0,0 +1,24 @@
+# pandas3.py
+import pandas as pd
+
+weekly_data = {'day':['Monday','Tuesday', 'Wednesday', 'Thursday',
+                      'Friday', 'Saturday', 'Sunday'],
+                'temp':[40, 33, 42, 31, 41, 40, 30],
+                'condition':['Sunny','Cloudy','Sunny','Rain','Sunny',
+                        'Cloudy','Rain']
+        }
+
+df = pd.DataFrame(weekly_data)
+df.index = ['MON', 'TUE','WED','THU','FRI','SAT','SUN']
+#Provide row with label TUE
+print(df.loc['TUE'])
+#Provide two rows with label TUE and WED
+print(df.loc[['TUE','WED']])
+#provide a temp value from row with label FRI
+print(df.loc['FRI','temp'])
+#Provide a row with index 2
+print(df.iloc[2])
+#provide a value from a location with
+# row index 2 and column index 2
+print(df.iloc[2,2])
+
@@ -0,0 +1,17 @@
+# pandas4.py
+import pandas as pd
+
+weekly_data = {'day':['Monday','Tuesday', 'Wednesday', 'Thursday',
+                      'Friday', 'Saturday', 'Sunday'],
+                'temp':[40, 33, 42, 31, 41, 40, 30],
+                'condition':['Sunny','Cloudy','Sunny','Rain','Sunny',
+                        'Cloudy','Rain']
+        }
+
+df = pd.DataFrame(weekly_data)
+df.index = ['MON', 'TUE','WED','THU','FRI','SAT','SUN']
+df.loc['TST1'] = ['Test day 1', 50, 'NA']
+df.loc[7] = ['Test day 2', 40, 'NA']
+
+print(df)
+
@@ -0,0 +1,27 @@
+# pandas5.py
+import pandas as pd
+
+weekly_data = {'day':['Monday','Tuesday', 'Wednesday', 'Thursday',
+                      'Friday', 'Saturday', 'Sunday'],
+                'temp':[40, 33, 42, 31, 41, 40, 30],
+                'condition':['Sunny','Cloudy','Sunny','Rain','Sunny',
+                        'Cloudy','Rain']
+        }
+
+df = pd.DataFrame(weekly_data)
+df.index = ['MON', 'TUE','WED','THU','FRI','SAT','SUN']
+
+#Adding a new column and then updating it
+df['Humidity1'] = [60, 70, 65,62,56,25,'']
+df['Humidity1'] = [60, 70, 65,62,56,251,'']
+
+#Inserting a colun at colum index of 2
+df.insert(2, "Humidity2",[60, 70, 65,62,56,25,''])
+#df.insert(2, "Humidity2",[60, 70, 65,62,56,25,''])
+
+#Adding two columns
+df1 = df.assign(Humidity3 = [60, 70, 65,62,56,25,''],Humidity4 = [60, 70, 65,62,56,25,''])
+print(df1)
+
+
+
@@ -0,0 +1,17 @@
+# pandas6.py
+import pandas as pd
+
+weekly_data = {'day':['Monday','Tuesday', 'Wednesday', 'Thursday',
+                      'Friday', 'Saturday', 'Sunday'],
+                'temp':[40, 33, 42, 31, 41, 40, 30],
+                'condition':['Sunny','Cloudy','Sunny','Rain','Sunny',
+                        'Cloudy','Rain']
+        }
+
+df = pd.DataFrame(weekly_data)
+df.index = ['MON', 'TUE','WED','THU','FRI','SAT','SAT']
+print(df)
+print(df.reset_index(drop=True))
+
+
+
@@ -0,0 +1,20 @@
+# pandas7.py
+import pandas as pd
+
+weekly_data = {'day':['Monday','Tuesday', 'Wednesday', 'Thursday',
+                      'Friday', 'Saturday', 'Sunday'],
+                'temp':[40, 33, 42, 31, 41, 40, 30],
+                'condition':['Sunny','Cloudy','Sunny','Rain','Sunny',
+                        'Cloudy','Rain']
+        }
+
+df = pd.DataFrame(weekly_data)
+df.index = ['MON', 'TUE','WED','THU','FRI','SAT','SUN']
+print(df)
+df1=df.drop(index=['SUN','SAT'])
+df2=df1.drop(columns=['condition'])
+
+print(df2)
+
+
+
@@ -0,0 +1,20 @@
+# pandas8.py
+import pandas as pd
+
+weekly_data = {'day':['Monday','Tuesday', 'Wednesday', 'Thursday',
+                      'Friday', 'Saturday', 'Sunday'],
+                'temp':[40, 33, 42, 31, 41, 40, 30],
+                'condition':['Sunny','Cloudy','Sunny','Rain','Sunny',
+                        'Cloudy','Rain']
+        }
+
+df = pd.DataFrame(weekly_data)
+df.index = ['MON', 'TUE','WED','THU','FRI','SAT','SUN']
+print(df)
+df1=df.rename(index={'SUN': 'SU', 'SAT': 'SA'})
+df2=df1.rename(columns={'condition':'cond'})
+
+print(df2)
+
+
+
@@ -0,0 +1,8 @@
+day,temperature,condition
+Monday,40,Sunny
+Tuesday,38,Sunny
+Wednesday,42,Sunny
+Thursday,37,Rain
+Friday,35,Cloudy
+Saturday,40,Cloudy
+Sunday,30,Rain