WEBVTT 00:01.220 --> 00:05.420 In this video you will learn when and why to use the copy method. 00:05.420 --> 00:10.460 So first of all we import the pandas and we import our Titanic data set. 00:10.460 --> 00:17.300 So that's the first five rows here and in the last videos we selected a numerical or a non numerical 00:17.300 --> 00:21.270 column and we analyzed the column and played around with it. 00:21.290 --> 00:28.610 That's a typical workflow with pandas selecting a subset of a data frame for example a column and playing 00:28.610 --> 00:29.590 around with it. 00:29.670 --> 00:36.800 But actually we do not want to change the original data frame and that sounds easy but there is a severe 00:36.800 --> 00:42.780 pitfall here that causes one of the most frequently made mistakes when working with pandas. 00:42.950 --> 00:49.700 Again we select here the H column and save for the H column and the variable h here and let's have a 00:49.700 --> 00:49.880 look. 00:49.880 --> 00:52.820 So these are the first five elements. 00:52.820 --> 00:55.370 And let's assume that we want to change. 00:55.460 --> 00:58.340 The third element here at the next position too. 00:58.730 --> 01:04.880 And the peasants of here is not the 26 year old but the 29 years old and we can actually change and 01:04.940 --> 01:08.450 overwrite that element by selecting the element here. 01:08.450 --> 01:13.760 For example with the ILO operator and we assign the new value twenty nine. 01:13.940 --> 01:16.000 So let's do this here. 01:16.700 --> 01:19.110 And here we get a warning message. 01:19.190 --> 01:22.610 This so-called setting with copy warning. 01:22.610 --> 01:28.130 And it says here a value is trying to be set on a copy of a slice from a data frame. 01:28.130 --> 01:31.270 So this sounds weird actually but let's simply have a look. 01:31.400 --> 01:33.230 What we did here with our code. 01:33.440 --> 01:36.890 So we have a look here first of all on our H column here. 01:37.040 --> 01:39.120 So that's the five elements. 01:39.290 --> 01:44.750 And we can see here that we changed to the third element from 26 to 29. 01:44.870 --> 01:51.830 And let's also look here at our Titanic data frame and the intention was actually not to change the 01:51.830 --> 01:57.340 element here in the overall data frame but that's exactly what we did here. 01:57.350 --> 02:02.960 So also here we have 29 and that's really a severe pitfall yet pandas. 02:03.080 --> 02:07.780 So we only intended to change one element in a slice of a data frame. 02:07.790 --> 02:09.280 So we in our each column. 02:09.500 --> 02:10.480 But the. 02:10.520 --> 02:17.930 We effectively changed also the overall data frame and there's a simple back around to prevent this 02:17.930 --> 02:19.180 behavior here. 02:19.370 --> 02:21.130 And let's go up again here. 02:21.830 --> 02:24.350 So let's rerun here the sets. 02:24.370 --> 02:26.300 So again we're in part here. 02:26.300 --> 02:28.880 The Titanic is yes we file. 02:28.880 --> 02:35.390 So again here we have 26 and now we want to select the H column and we want to play around with the 02:35.390 --> 02:41.110 H column but we do not want to change to the original data frame and that's a simple background. 02:41.120 --> 02:43.730 So we are changing the copy method 02:47.300 --> 02:50.270 and by doing so pen creates a separate object. 02:50.270 --> 02:55.180 So our each column that is stored in the variable H is completely independent. 02:55.190 --> 03:02.290 Now from our Titanic data frame and before without the copy method the object that was stored in the 03:02.290 --> 03:08.720 verbal H was not independent from the data frame and in fact the object that was stored in the available 03:08.720 --> 03:16.910 H was a view on the data frame so it was actually still part of the overall data frame and we can prevent 03:16.910 --> 03:21.350 this by creating a copy here and let's do this here 03:24.440 --> 03:27.500 and let's repeat the process that we changed here. 03:27.500 --> 03:30.700 The third element 229. 03:30.850 --> 03:36.650 And here we can see we do not get any warning message and it's also checking the consequences on the 03:36.650 --> 03:40.070 object that is start here in the variable H. 03:40.100 --> 03:43.390 So here again we have changed to the third element. 03:43.580 --> 03:47.730 But know our original data frame we did not change the elements. 03:47.730 --> 03:50.350 So here we have still 26. 03:50.450 --> 03:53.720 So that's the most important use case for the copy method. 03:53.720 --> 03:57.410 And with this we are finished here and I hope to see us on the next video by.