WEBVTT 00:00.760 --> 00:07.330 In the last video we imported our very first time series data and we converted the column daytime into 00:07.330 --> 00:10.090 the daytime index of our data frame. 00:10.240 --> 00:16.970 And we did this by passing the column labor to the pyramids past states and index call of the Read CSP 00:17.020 --> 00:18.000 method. 00:18.160 --> 00:23.680 And in this video we will have a look at the second more explicit alternative to do this and two we 00:23.680 --> 00:24.460 need two steps. 00:24.460 --> 00:29.920 So first of all we transform the data type of the column to date time and we can do this. 00:29.920 --> 00:34.310 This is the direct pen us method to date time and the second. 00:34.330 --> 00:41.320 We said the column is the index of our data frame this the asset index method and then in the second 00:41.320 --> 00:48.160 part of this video we will have a closer look to the to date time method and we plan in more detail 00:48.160 --> 00:52.070 how to transform different strings into daytime objects. 00:52.270 --> 01:02.470 So let's import pandas here and we import our temperate CSP file and TV for the first five rows and 01:02.470 --> 01:04.920 we also quality of the Inform method. 01:04.930 --> 01:09.040 So here we have now the column day time and we offer the data type object. 01:09.040 --> 01:15.100 So these are strings here and we can also explicitly select the date time column. 01:15.100 --> 01:16.020 So let's do this 01:18.740 --> 01:26.090 and we can see at the end of that year we have the data type object and we can also select a single 01:26.090 --> 01:30.810 element of our panda series or column slow. 01:30.820 --> 01:36.950 Here we can see that this is actually a string and now we can convert the data type of our column to 01:36.950 --> 01:41.440 date time with the direct pen us method to date time. 01:41.450 --> 01:44.420 So here we pass our column date time 01:47.350 --> 01:51.850 and also you will get a pen that serious with all of our dates here. 01:51.990 --> 01:57.430 And if you have a look here at the very end we can see here that we have to date that type of daytime 01:57.430 --> 02:00.170 60 for service of the method daytime. 02:00.230 --> 02:05.970 We can transform strings into a daytime objects and now we're going to want to have the daytime column 02:05.970 --> 02:12.900 here as being the daytime index of our whole data frame and we can do this in two steps. 02:12.900 --> 02:18.570 So first of all we convert the data type with the method to data time and by doing so we're getting 02:18.570 --> 02:25.280 you actually hear the pandas series and then we pass here the pandas serious to the method the set index. 02:25.350 --> 02:32.800 So let's try this out here so we can see them now on the left hand side that instead of having a range 02:32.830 --> 02:40.300 index we have our column date time and still we have the column day time so we can drop the column here 02:43.000 --> 02:44.830 and now this looks quite good. 02:44.830 --> 02:49.380 And uh then we can override our variable temperature so let's do this 02:52.840 --> 02:59.340 and that's again never look at the first five rows so here we have our new data frame we have on the 02:59.340 --> 03:05.160 left hand side our daytime index and we have the two columns L.A. and New York. 03:05.160 --> 03:12.330 And let's also inspect the info method so we can see now our daytime indexes are worth thirty five thousand 03:12.330 --> 03:17.850 entries and our two columns and we can also further inspect our daytime index. 03:17.870 --> 03:19.600 But the attribute index. 03:19.620 --> 03:25.680 So here we have the daytime index with all of the time stems and we can also select a single element 03:25.680 --> 03:26.410 here. 03:26.430 --> 03:33.210 So the first one and here we have a timestamp and actually panned US is quite smart and versatile when 03:33.210 --> 03:36.260 it comes to converting strings to daytime optics. 03:36.390 --> 03:39.350 So actually panels us can handle many different formats. 03:39.390 --> 03:45.000 And uh the most send out format is actually a 215 or 520. 03:45.030 --> 03:50.990 So we have here the 20th of May 2015 and we have here the format year. 03:51.000 --> 03:52.710 Then we have a minus or dash. 03:52.710 --> 03:57.250 Then we have months then and then minus or dash and then we have the day. 03:57.300 --> 04:01.230 And this is no problem to convert into a daytime object. 04:01.350 --> 04:03.090 So here we get the timestamp. 04:03.390 --> 04:09.360 So you know our string we have the information about the date but uh we haven't passed any information 04:09.360 --> 04:10.390 on the time. 04:10.440 --> 04:13.470 And by default US assumes actually at midnight. 04:13.490 --> 04:16.400 So we have uh Midnight Zero clock actually. 04:16.420 --> 04:23.190 So this is here for the hours the minutes and the seconds but within I was doing we can also pass information 04:23.190 --> 04:24.530 on the time. 04:24.600 --> 04:32.820 So let's do this for example 10:00 and then we get to your 10:00 and we can also provide information 04:32.820 --> 04:33.690 on the minutes. 04:33.690 --> 04:36.450 So let's assume 10 30. 04:38.280 --> 04:41.910 And we can also provide additional information on these seconds. 04:41.910 --> 04:44.470 So for example 20 seconds. 04:44.610 --> 04:51.300 And as I said before I actually PANDAS and the method to date time is quite powerful and versatile because 04:51.300 --> 04:55.230 it can actually recognize and convert many different formats. 04:55.230 --> 04:57.770 So for example here we have a different format. 04:57.810 --> 05:02.480 So we have between year month and day no separate at all. 05:02.510 --> 05:04.260 But this still works. 05:04.260 --> 05:06.630 And again here by default we have at midnight. 05:06.690 --> 05:12.480 If we do not pass any information on the time and we can also use different separate us. 05:12.480 --> 05:14.670 So for example here slashes 05:17.420 --> 05:24.620 or we can use that byte spaces and that's only one thing that we have to keep in mind here. 05:24.630 --> 05:29.940 So by default the sequence this year month and day. 05:30.300 --> 05:33.780 And for example if you pass now a year day and month. 05:33.780 --> 05:39.460 So our intention here is to get the data the 20th of May 2015. 05:39.480 --> 05:44.870 But here we switched the day and month and here penned us gets the problem. 05:44.880 --> 05:51.310 So here we get an error message and it says months must be between one and twelve. 05:51.360 --> 05:58.710 So by default pen US expects here the sequence that year month and day and this is definitely a source 05:58.710 --> 06:04.860 for errors but we can also prevent this by having for the month for example in May. 06:05.430 --> 06:14.160 So this also works and we can also locate the May at the very left side of our string or on the right 06:14.160 --> 06:15.800 hand side. 06:15.990 --> 06:21.390 And actually it doesn't matter if we have here uppercase let us all lowercase letters or whatever. 06:21.480 --> 06:22.540 So it works. 06:22.680 --> 06:27.870 And instead of having 20 for the day become can also for your 20th with the T H. 06:27.870 --> 06:29.560 So this also works. 06:29.580 --> 06:33.000 So to summarize so there's hardly any combination here. 06:33.020 --> 06:40.650 It does not work and we can also pass a sequence with dates to the to date time method. 06:40.650 --> 06:47.530 So for example here we have a list and here above we passed a panda series. 06:48.440 --> 06:50.310 Our column dates time 06:53.070 --> 06:56.180 and now we simply pass a list with two dates. 06:56.220 --> 07:03.960 And this also works here and whenever we pass more than one date we get here a date time index. 07:03.990 --> 07:12.030 So here you have now a day time index with two time stamps and we can also pass a sequence of dates 07:12.180 --> 07:14.220 where for at least one element. 07:14.490 --> 07:19.590 Even pandas will not be able to convert this into daytime so for example elephant. 07:19.800 --> 07:26.880 So this might be impossible to convert but we can try here and let's have a look what daytime does here 07:27.540 --> 07:34.620 and by default we get here an error message so unknown string format elephant. 07:34.620 --> 07:38.170 So we have senior buffer that pandas can handle many string formats. 07:38.190 --> 07:42.060 But of course elephant is quite impossible to convert. 07:42.240 --> 07:44.190 And by default we are getting an error message. 07:44.190 --> 07:46.050 But we can change this. 07:46.050 --> 07:52.090 So let's have a look here and here we have the parameter arrows. 07:52.370 --> 07:54.360 And by default it's set to raise. 07:54.410 --> 08:03.220 So let's have a look here at the documentation so if we have the argument raise then the invalid passing 08:03.220 --> 08:04.420 that raise an exception. 08:04.420 --> 08:10.150 So we're getting an error message but we can also possibly coerce and then the invalid passing will 08:10.150 --> 08:12.010 be set as an 80 year. 08:12.070 --> 08:14.050 So this stands for not a time. 08:14.770 --> 08:18.940 So let's try this out and let's pass codes to the parameter arrows 08:23.370 --> 08:30.240 and it's from the cell here and here we can see that we nevertheless get a daytime index with the two 08:30.240 --> 08:37.410 time stamps and actually the element an 80 year which stands for not a time. 08:37.410 --> 08:44.490 So also here and having some strings which cannot be converted to daytime then still the daytime method 08:44.550 --> 08:46.620 is quite smart and versatile. 08:47.340 --> 08:47.910 All right. 08:47.940 --> 08:52.590 With this we are finished now with this video and I hope to see you also in the next one by.