WEBVTT 00:01.020 --> 00:07.710 In this video we learn how to handle and deal with a mighty index that we create and group by operations. 00:07.710 --> 00:13.380 So when performing group eye operations in many cases we are getting a data frame with a mighty index 00:13.740 --> 00:16.300 also called hierarchical index. 00:16.380 --> 00:23.160 So let's have a look at the example and first of all we are importing pandas and we are working with 00:23.160 --> 00:31.440 the Titanic data set and we only import these survived PE class six eights and Fair column. 00:31.440 --> 00:39.920 So let's have a look at our data frame so here we have the five columns and now we want to group our 00:39.920 --> 00:43.250 data frame by the sixth column and peak last column. 00:43.460 --> 00:49.700 So we are actually creating six groups and for each group we actually want to calculate the survival 00:49.700 --> 00:50.520 rate. 00:50.540 --> 00:58.130 The average age and the average fan we can do this by applying the mean method in our group by object. 00:58.130 --> 01:03.680 So nothing new here and we are starting the residing data frame and the average summary. 01:03.890 --> 01:09.260 So let's do this and let's have a look at our new data frame summary. 01:12.300 --> 01:19.240 So here we can see we have the survival rate that the average age and the average fare for all six groups. 01:19.260 --> 01:25.020 So for example we have the group female passengers in the first class female passengers and the second 01:25.020 --> 01:31.350 class and so on and here on the left hand side we actually see a mighty index. 01:31.410 --> 01:38.640 So to say we've actually had two columns that serve s the index and we fear the outer index level. 01:38.670 --> 01:47.420 But to see the sex column and we have the index level PE class and our mighty index data frame is here 01:47.420 --> 01:50.050 first of all sorted by the Alta index. 01:50.040 --> 01:55.800 So first of all we have here the female passengers and then we have 50 male passengers. 01:55.800 --> 01:58.020 And within the groups female and male. 01:58.380 --> 02:05.010 So the OCR sorted by the peasants class so first second and third first second and third. 02:05.010 --> 02:11.340 So by creating your a group by object by more than one column and applying the mean method we are actually 02:11.340 --> 02:15.530 creating here a brand new data frame with a mighty index. 02:15.720 --> 02:22.950 So the outer index level is the sex column and the index level is the peak class column and we can also 02:22.950 --> 02:27.530 check the index type with the index attribute. 02:27.570 --> 02:35.550 So here we can see we have a mighty index and also with a mighty index we can perform label based indexing 02:35.850 --> 02:36.930 and we can do this here. 02:37.000 --> 02:42.000 At the local operator and we are opening here a scrap record. 02:42.050 --> 02:45.700 Now let's assume that we only want to select this row here. 02:45.810 --> 02:53.190 So females in the second class and we want to f your column so survived H and fare and we can do this 02:53.190 --> 02:56.850 by passing year on the right hand side of the comma. 02:57.070 --> 02:59.430 The column labels that we want to have here. 02:59.440 --> 03:07.170 So in this example all columns and on the left hand side of the comma we have to slice our mighty index 03:07.170 --> 03:12.670 and we can do this with the brackets here and again here we have a comma separator. 03:12.690 --> 03:19.140 And on the left hand side of the comma be defined which labels of our outer index we want to f. 03:19.140 --> 03:26.430 So your few methods and on the right hand side of the comma you can define which labels of our in the 03:26.460 --> 03:27.690 index we want to f. 03:27.720 --> 03:36.220 So you're a second class so let's do this a. getting actually all information off this row. 03:36.490 --> 03:44.020 So we have female and the second class and we have a survival rate of 92 percent and we have an average 03:44.050 --> 03:52.640 age of twenty eight years and we have an average ticket price of twenty one point nine and we can also 03:52.640 --> 03:54.020 select a single column. 03:54.020 --> 04:00.110 So instead of selecting all the columns that can pass here for example the column labeled H 04:03.380 --> 04:09.580 A.B. get to the Everett's h of all female peasants us in the second class or twenty eight point seven 04:09.590 --> 04:18.160 yes and with the slope level method we can actually swap here our index levels so we can swap our inner 04:18.160 --> 04:26.200 index PE class to the outer index level and we can swap our outer in Next Level Six to the next level. 04:26.530 --> 04:32.100 So let's try this out here and let's cross out here to sort in next so we'll do this later. 04:36.120 --> 04:42.240 So here we can see that now on the left hand side as the altar and next level we have peak class and 04:42.240 --> 04:48.750 on the right hand side as the inner index level we have sex and every time and we are creating a new 04:48.750 --> 04:49.120 model. 04:49.130 --> 04:54.000 And next that definitely makes sense to sort our index with the assault index method 04:57.360 --> 05:03.660 so here we can see we have our sorted mighty index and we have first of all the first class then the 05:03.660 --> 05:08.330 second class and the third class and within each class we have the order. 05:08.330 --> 05:09.980 Female and male. 05:10.680 --> 05:16.500 And if we are not happy with a mighty index we can also drop the mighty index and we can create an ordinary 05:16.500 --> 05:18.390 Range Index and we can do this. 05:18.390 --> 05:20.880 This the the reset index method 05:24.820 --> 05:29.040 so here we can see we dropped our mighty and next and we have a simple range. 05:29.080 --> 05:37.660 And next year starting from 0 to 5 and here we have sex and PE class as now ordinary columns of our 05:37.660 --> 05:38.830 new data frame here. 05:39.220 --> 05:44.320 So actually when we are performing group by operations in many cases we are creating a data frame with 05:44.320 --> 05:45.390 a mighty index. 05:45.670 --> 05:51.400 But if we are not happy with a mighty index we can simply drop the mighty index this. 05:51.460 --> 05:57.250 The reset index method and also here we can index and slice here for different values. 05:57.250 --> 05:58.940 So this is no problem. 05:58.960 --> 06:05.620 So there's actually no need to work with data frames with multi indexes but sometimes it makes our life 06:05.650 --> 06:13.390 easier and we should definitely know how to handle and work with the data frames that have multi indexes. 06:13.390 --> 06:17.680 All right so now we are finished with this video and I hope to see in the next video by.