1
00:00:00,005 --> 00:00:02,009
- [Instructor] Data frames make it easy to perform

2
00:00:02,009 --> 00:00:06,004
all kinds of manipulation on the data that they contain.

3
00:00:06,004 --> 00:00:07,005
So for this example,

4
00:00:07,005 --> 00:00:11,004
I'm going to focus on several of the more common operations.

5
00:00:11,004 --> 00:00:15,000
Let's open up pandas_manipulate.

6
00:00:15,000 --> 00:00:17,000
And you can see in my example file,

7
00:00:17,000 --> 00:00:21,006
I already have some code that reads the Inventory.csv file

8
00:00:21,006 --> 00:00:24,000
into a data frame.

9
00:00:24,000 --> 00:00:28,000
So for the first example, let's create a new column of data.

10
00:00:28,000 --> 00:00:30,004
We did this earlier in the course manually,

11
00:00:30,004 --> 00:00:31,006
but now we're going to accomplish

12
00:00:31,006 --> 00:00:33,009
the same thing using pandas.

13
00:00:33,009 --> 00:00:36,008
We're going to add a column named margin

14
00:00:36,008 --> 00:00:39,005
that represents the difference between the consumer price

15
00:00:39,005 --> 00:00:42,001
and the wholesale price for each product.

16
00:00:42,001 --> 00:00:45,002
So all I need to do is name the new column

17
00:00:45,002 --> 00:00:47,006
and use a formula to calculate the difference

18
00:00:47,006 --> 00:00:49,004
between the other two columns.

19
00:00:49,004 --> 00:00:51,006
So watch how easy this is.

20
00:00:51,006 --> 00:00:54,008
I'm just simply going to on the data frame make a new column.

21
00:00:54,008 --> 00:00:58,006
I'm going to call that margin.

22
00:00:58,006 --> 00:01:01,006
And I'm going to set that equal to the difference

23
00:01:01,006 --> 00:01:07,007
between the consumer price.

24
00:01:07,007 --> 00:01:10,001
And I'm going to subtract off

25
00:01:10,001 --> 00:01:15,003
the column for the wholesale price.

26
00:01:15,003 --> 00:01:16,007
And that's all there is to it.

27
00:01:16,007 --> 00:01:19,001
This will create a new column named margin

28
00:01:19,001 --> 00:01:23,004
and calculate the difference for each row in the data set.

29
00:01:23,004 --> 00:01:26,000
And let's also print out the shape of the data

30
00:01:26,000 --> 00:01:39,000
along with the first five rows.

31
00:01:39,000 --> 00:01:43,008
All right, so let's go ahead and save our code

32
00:01:43,008 --> 00:01:51,009
and let's run this.

33
00:01:51,009 --> 00:01:54,005
All right, and you can see that now we have a new column.

34
00:01:54,005 --> 00:01:58,002
So the shape has changed from being 50 columns,

35
00:01:58,002 --> 00:02:00,007
or, I'm sorry, 50 rows and five columns

36
00:02:00,007 --> 00:02:03,002
to 50 rows and six columns.

37
00:02:03,002 --> 00:02:05,007
And we can see that the new margin column

38
00:02:05,007 --> 00:02:08,000
is here in the result.

39
00:02:08,000 --> 00:02:13,000
So we've modified the shape of our data structure.

40
00:02:13,000 --> 00:02:14,008
We can also modify the contents

41
00:02:14,008 --> 00:02:17,006
of a column directly in place.

42
00:02:17,006 --> 00:02:19,006
Now, there's a variety of ways to do this,

43
00:02:19,006 --> 00:02:21,007
and the method I'm going to use

44
00:02:21,007 --> 00:02:24,005
is by using the apply function.

45
00:02:24,005 --> 00:02:28,007
So the apply function allows you to apply another function

46
00:02:28,007 --> 00:02:31,005
to each cell in a particular column.

47
00:02:31,005 --> 00:02:34,004
So suppose I wanted to modify the cells

48
00:02:34,004 --> 00:02:36,007
in the category column.

49
00:02:36,007 --> 00:02:39,000
Let's create some code to do that.

50
00:02:39,000 --> 00:02:42,002
And I'm going to comment out the previous example.

51
00:02:42,002 --> 00:02:44,007
So on the data frame,

52
00:02:44,007 --> 00:02:49,009
I'm going to specify I want to work on the category column.

53
00:02:49,009 --> 00:02:54,003
And what I'm going to do is assign that to the result

54
00:02:54,003 --> 00:02:59,005
of the apply function.

55
00:02:59,005 --> 00:03:03,000
Now, the argument to the apply function is any function

56
00:03:03,000 --> 00:03:07,003
I want to call for each cell in the column.

57
00:03:07,003 --> 00:03:09,002
So let's just make this simple.

58
00:03:09,002 --> 00:03:11,003
Let's convert each string

59
00:03:11,003 --> 00:03:14,000
in the category column to uppercase.

60
00:03:14,000 --> 00:03:16,006
To do that, I'm going to just supply

61
00:03:16,006 --> 00:03:21,006
an inline lambda function for that.

62
00:03:21,006 --> 00:03:24,003
And I'm going to call x.upper.

63
00:03:24,003 --> 00:03:26,008
So the argument to my lambda function,

64
00:03:26,008 --> 00:03:27,009
there's going to be one argument.

65
00:03:27,009 --> 00:03:29,004
It's going to be the string,

66
00:03:29,004 --> 00:03:32,000
which is the value of each cell in the column.

67
00:03:32,000 --> 00:03:33,001
And since it's a string,

68
00:03:33,001 --> 00:03:35,000
I can use the python upper function

69
00:03:35,000 --> 00:03:37,003
to convert it to uppercase.

70
00:03:37,003 --> 00:03:39,001
And let's also print the data frame

71
00:03:39,001 --> 00:03:43,000
when we're done with that.

72
00:03:43,000 --> 00:03:45,006
All right, so let's go ahead and save.

73
00:03:45,006 --> 00:03:50,000
And now let's run.

74
00:03:50,000 --> 00:03:54,009
And here you can see in the result in my category column,

75
00:03:54,009 --> 00:03:59,009
all of the strings have now been converted to uppercase.

76
00:03:59,009 --> 00:04:02,001
If you want to rename the columns on the data frame,

77
00:04:02,001 --> 00:04:04,002
that's also pretty easy.

78
00:04:04,002 --> 00:04:06,005
We can use the rename function for this.

79
00:04:06,005 --> 00:04:09,000
And again, there are a few ways to call this function,

80
00:04:09,000 --> 00:04:10,007
depending on your use case,

81
00:04:10,007 --> 00:04:13,008
but I'll demonstrate how to rename specific columns.

82
00:04:13,008 --> 00:04:17,003
So let's assume we want to rename wholesale price

83
00:04:17,003 --> 00:04:20,009
and consumer price to just wholesale and consumer.

84
00:04:20,009 --> 00:04:22,008
We can use the rename function

85
00:04:22,008 --> 00:04:25,000
and pass a dictionary of old

86
00:04:25,000 --> 00:04:28,002
and new names using the columns parameter.

87
00:04:28,002 --> 00:04:29,004
So on the data frame,

88
00:04:29,004 --> 00:04:33,007
I'm going to call the rename function,

89
00:04:33,007 --> 00:04:37,002
and I'm going to specify that I'm renaming columns.

90
00:04:37,002 --> 00:04:38,005
And inside this object,

91
00:04:38,005 --> 00:04:42,002
I'm going to give key value pairs

92
00:04:42,002 --> 00:04:47,006
and the key is going to be the column I want to be renamed

93
00:04:47,006 --> 00:04:51,005
paired with what the new name is.

94
00:04:51,005 --> 00:04:54,002
So let's do that for wholesale price,

95
00:04:54,002 --> 00:04:58,005
and let's also do consumer price.

96
00:04:58,005 --> 00:05:04,009
We will rename that to just consumer.

97
00:05:04,009 --> 00:05:06,006
And then I have to specify that I want this

98
00:05:06,006 --> 00:05:08,004
to happen in place.

99
00:05:08,004 --> 00:05:12,008
So I'm going to set the in place argument to be true.

100
00:05:12,008 --> 00:05:14,006
And then once that's done,

101
00:05:14,006 --> 00:05:15,009
we'll just call the head function

102
00:05:15,009 --> 00:05:20,003
to print out the first five lines.

103
00:05:20,003 --> 00:05:23,009
Let me comment out the previous example.

104
00:05:23,009 --> 00:05:31,009
All right, so let's go ahead and save this and let's run it.

105
00:05:31,009 --> 00:05:37,008
And oh, whoop, I have to actually print the head,

106
00:05:37,008 --> 00:05:40,006
not just call it.

107
00:05:40,006 --> 00:05:46,006
All right, let's try that again.

108
00:05:46,006 --> 00:05:50,006
All right, and so now we can see that the wholesale price

109
00:05:50,006 --> 00:05:54,007
and consumer price columns have been properly renamed.

110
00:05:54,007 --> 00:05:55,005
All right, one more thing.

111
00:05:55,005 --> 00:05:59,001
Let's finally look at how to delete a column of data.

112
00:05:59,001 --> 00:06:00,007
So to delete a column of data,

113
00:06:00,007 --> 00:06:02,007
we use the drop function

114
00:06:02,007 --> 00:06:06,002
with the name of the column that we no longer want.

115
00:06:06,002 --> 00:06:10,005
So let's comment on this.

116
00:06:10,005 --> 00:06:13,008
So on the data frame, I'm going to use the drop function.

117
00:06:13,008 --> 00:06:17,008
So let's drop the margin column that we've already made.

118
00:06:17,008 --> 00:06:19,000
And in order to do that,

119
00:06:19,000 --> 00:06:20,008
I'll have to re-enable the code

120
00:06:20,008 --> 00:06:26,008
that creates the margin column.

121
00:06:26,008 --> 00:06:29,000
All right, and in the drop function,

122
00:06:29,000 --> 00:06:30,004
I have to specify, once again,

123
00:06:30,004 --> 00:06:33,007
I want this to happen in place.

124
00:06:33,007 --> 00:06:35,003
I don't want to create a new data frame.

125
00:06:35,003 --> 00:06:37,003
So I'm going to set in place equals to true.

126
00:06:37,003 --> 00:06:40,003
And then I have to specify what's called the axis.

127
00:06:40,003 --> 00:06:42,006
The axis I'm going to set to the value of one.

128
00:06:42,006 --> 00:06:44,009
This basically means to operate on the columns,

129
00:06:44,009 --> 00:06:46,001
not the rows,

130
00:06:46,001 --> 00:06:48,006
because otherwise the pandas library doesn't know

131
00:06:48,006 --> 00:06:51,003
whether this is a row index label

132
00:06:51,003 --> 00:06:52,007
or a column index label.

133
00:06:52,007 --> 00:06:56,005
So by specifying the axis parameter with a value of one,

134
00:06:56,005 --> 00:06:58,001
I'm saying, hey, this is a column,

135
00:06:58,001 --> 00:07:02,006
so go look for the column named margin and drop that.

136
00:07:02,006 --> 00:07:03,009
And then after we've done that,

137
00:07:03,009 --> 00:07:10,005
let's go ahead and print out the head one more time.

138
00:07:10,005 --> 00:07:15,001
So let's re-enable the code that shows the margin there,

139
00:07:15,001 --> 00:07:18,006
and then we'll print it again after it's gone.

140
00:07:18,006 --> 00:07:26,003
So let's go ahead and save this and let's try it.

141
00:07:26,003 --> 00:07:27,005
All right, so in the first example,

142
00:07:27,005 --> 00:07:30,004
we can see that we created the margin column properly

143
00:07:30,004 --> 00:07:33,008
and then we can see that it's been dropped.

144
00:07:33,008 --> 00:07:35,001
All right, so this should give you some sense

145
00:07:35,001 --> 00:07:36,001
of how powerful

146
00:07:36,001 --> 00:07:39,004
and flexible data frames are for working with data content.

147
00:07:39,004 --> 00:07:41,004
And I would suggest maybe spending some time here

148
00:07:41,004 --> 00:07:42,006
referring to the docs

149
00:07:42,006 --> 00:07:44,009
and trying out some of your own experiments

150
00:07:44,009 --> 00:07:46,000
before moving on.