WEBVTT 00:01.210 --> 00:06.740 In the last couple of videos we have learned how to transform a sequence for example a panda series 00:07.220 --> 00:13.410 or a list containing strings with the date time information into a date time index. 00:13.430 --> 00:20.030 So let's import pandas and let's assume that here we have a list with two strings. 00:20.030 --> 00:28.610 So the 20th of May to 15 and the 20th of February to 15 and we can pass here the list to the pandas 00:28.610 --> 00:36.850 method to date time and by doing so we are getting here a day time index with two timestamps. 00:37.020 --> 00:41.920 Now let's assume that we want to create a daytime index service many time stamps. 00:41.930 --> 00:49.760 So for example if we want to f all 31 days in July to 15 so creating and passing all 31 strings to the 00:49.810 --> 00:54.090 two day time method many earlier is actually not really efficient. 00:54.230 --> 01:00.140 And if we have constant frequency so constant increments between the time stamps then we can simply 01:00.140 --> 01:09.140 do this limit the direct Panda's method date range and in our example so all days in July to and fulfills 01:09.150 --> 01:10.370 them the condition. 01:10.460 --> 01:15.480 So we have here a daily frequency or increments of one day. 01:15.800 --> 01:21.460 So let's use here the direct pandas method date range and let's have a look here. 01:23.080 --> 01:29.750 So we have here the stop parameter and two we can pass the string with the date that we want to have 01:29.780 --> 01:30.930 a start. 01:30.940 --> 01:38.260 So here are the first of July and we have also the repair meter end and here we pass the thirty first 01:38.260 --> 01:38.980 off July. 01:39.370 --> 01:43.150 And then there's the additional parameter freq for frequency. 01:43.150 --> 01:45.610 And let's have a look here and here. 01:45.610 --> 01:48.130 The default setting is capital D. 01:48.160 --> 01:49.850 So this stands for days. 01:49.960 --> 01:53.670 So we can also explicitly set the frequency power meter. 01:53.670 --> 02:03.500 Two days this is just here for clarification and by doing so we're actually creating all days starting 02:03.500 --> 02:07.850 from the 1st of July onto the thirty first of July. 02:07.880 --> 02:15.310 So let's do this here and here again see we get to the first or the second the third onto the thirty 02:15.310 --> 02:16.190 first. 02:16.420 --> 02:24.130 And this is set a day time index with uh thirty one time stamps and the data types the date time 64 02:24.130 --> 02:31.600 data type and we have a frequency of D or so daily frequency and let's again have a look here at the 02:31.600 --> 02:34.110 documentation and that the pair meet us. 02:34.480 --> 02:38.200 So essentially we have three parameters we have the stock power meter. 02:38.410 --> 02:42.190 We have for the Empire meter and we have the periods power meter. 02:42.670 --> 02:49.700 And actually we have to define two out of these three parameters to call our date range method. 02:49.720 --> 02:58.180 And of course we have also to define the frequency so instead of passing the first two stop and the 02:58.210 --> 03:05.800 30 first to end B can also pass here the first to start and we can say we want to f thirty one periods 03:06.430 --> 03:09.040 and we want to have a daily frequency. 03:09.520 --> 03:17.500 And by doing so we are getting actually the same identical daytime index from the 1st of July onto the 03:17.530 --> 03:21.990 thirty first of July and s also a third alternative. 03:22.000 --> 03:29.770 So we could alternatively set here the end power meter and the periods pair meter so the end date is 03:29.820 --> 03:34.630 uh the thirty first of July and you want to f 31 periods. 03:34.900 --> 03:37.000 And again the frequency is daily. 03:38.680 --> 03:46.560 And also here we get the same identical daytime index so again we have three paramedics we have stop 03:46.650 --> 03:55.140 and and periods and we have to set are determined to out of these three parameters and additionally 03:55.140 --> 03:58.850 the frequency parameter and for the frequency parameter. 03:58.860 --> 04:06.750 We have many different options so feel free to have a look at the documentation to see all options and 04:06.750 --> 04:10.830 one option is to pass a B for business day. 04:10.860 --> 04:18.350 So again we have here the first of July a start and the thirty first up to my ass end and the U.S.A 04:18.390 --> 04:19.350 business days 04:22.660 --> 04:28.180 and we get the first the second the third and the sixth of July. 04:28.260 --> 04:33.810 So we do not have the fourth and the fifth because of this Saturday and Sunday. 04:34.260 --> 04:40.560 So Saturday and Sunday are not business days and therefore here in this day time index we have only 04:40.560 --> 04:43.580 business days or actually weekdays. 04:43.650 --> 04:46.460 So from Monday to Friday. 04:46.470 --> 04:46.890 All right. 04:46.890 --> 04:51.080 Now let's assume that our start date is the thirty first of July. 04:51.450 --> 04:59.490 And you want to have 10 timestamps with hourly frequency so we pass your h for hourly frequency 05:03.900 --> 05:09.190 and by default we are starting here at midnight on the thirty first of July. 05:09.690 --> 05:16.620 And then we have four year increments of one hour or so we have 1:00 2:00 3:00 until 9:00. 05:16.620 --> 05:22.980 So in total 10 timestamps because they pass the 10 to the parameter periods. 05:22.980 --> 05:31.960 So we have a daytime index that's 10 times stamps and the frequency is h for hourly. 05:32.050 --> 05:34.750 All right and we can also have a weekly frequency. 05:35.200 --> 05:38.380 So yeah our start date as the first of July. 05:38.800 --> 05:45.970 And we want to have six periods and you want to have weekly frequency so we pass yet w to the frequency 05:46.000 --> 05:46.720 parameter 05:49.430 --> 05:53.410 and now at a very first glance this looks quite strange. 05:53.420 --> 06:00.710 So our very first timestamp as of the 5th of July and not to the first of July and this is because by 06:00.710 --> 06:07.460 default if you pass your weekly clear then we get here only some days and actually penned us checks 06:07.460 --> 06:12.700 here from the Stop day bad the very first Sunday is and the very first Sunday is on the 5th. 06:13.400 --> 06:16.760 So here the 5th is our first time stamp. 06:16.750 --> 06:24.020 Then we have our second time stamp at the tomatoes and so on and the very last time stamp and the sixth 06:24.020 --> 06:27.680 time stamp is the 9th of August. 06:27.710 --> 06:35.780 So here we have a daytime index with six some days and we can see this here at the frequency we have 06:35.800 --> 06:43.300 here weekly Sunday and now let's assume that you want to have a starting from the 1st of July it only 06:43.300 --> 06:43.900 Wednesday. 06:43.930 --> 06:52.210 So we want to have the next six best days and we can do this by not passing the weekly but we clear 06:52.390 --> 06:53.870 then minus or dash. 06:53.920 --> 06:55.150 And then Wednesday 06:58.580 --> 07:01.850 and apparently the 1st of July is a Wednesday. 07:02.180 --> 07:04.970 So pandas starts from here. 07:04.970 --> 07:11.430 The starting point and then searches for the very first the Wednesday here in this case it is the 1st 07:11.450 --> 07:12.260 of July. 07:12.500 --> 07:17.690 And then here in total we have the next six Wednesdays from the 1st of July here. 07:17.900 --> 07:21.060 And actually we can create you all week day so we can pass here. 07:21.080 --> 07:27.890 If our AI for Friday are we weekend pass Tuesday and so on and here and the frequency we can see here 07:27.890 --> 07:28.360 that. 07:28.370 --> 07:30.650 Here we have Wednesdays. 07:30.650 --> 07:35.760 All right so now let's go to monthly frequency and for monthly we can pass here. 07:35.770 --> 07:42.590 M So our starting day is the 14th of July and we want to have six timestamps. 07:42.590 --> 07:45.730 So let's try this out here. 07:46.100 --> 07:52.340 And by default when we are passing the M to the frequency parameter we are getting it actually the month 07:52.430 --> 07:53.550 and dates. 07:53.600 --> 07:57.300 So pen that starts here at the 14th of July. 07:57.710 --> 08:04.520 And searches are the very first month and the date and this is the thirty first of July. 08:05.240 --> 08:10.420 And then we have the next consecutive five month ends. 08:11.000 --> 08:16.220 So we have for the thirty first of August the 30th of September and so on. 08:16.970 --> 08:22.790 And let's assume that instead of having the month and we want to for the very first day of each month 08:23.240 --> 08:28.410 and we can do this by passing here m s so far start. 08:28.640 --> 08:36.790 So we get the start of each month and again your partner starts here at the starting point and searches 08:36.790 --> 08:39.870 for the very first uh beginning of months. 08:39.880 --> 08:47.800 And this is here the first of August and then we FBA the first of September and so on and here we can 08:47.800 --> 08:51.350 see the frequency as the month start. 08:51.370 --> 08:51.880 All right. 08:51.880 --> 09:00.490 Now let's assume that we want to have the 14th of July August September October November and December. 09:00.490 --> 09:04.340 So we want to have six period starting from the 14th of July. 09:04.630 --> 09:12.130 And instead of passing your monthly or month start we can pass here a special pants method called Date 09:12.220 --> 09:19.870 offset and tier but then the brackets have to define that the frequency should be month slow the frequency 09:19.870 --> 09:20.850 is one month. 09:21.400 --> 09:23.530 So let's try this out here. 09:23.650 --> 09:31.680 And by doing so we're creating the timestamps 14th of July 14th of August and so on and we can see in 09:31.690 --> 09:37.240 the frequency that we have a date offset of month equals 1. 09:37.240 --> 09:40.650 All right so now let's head to quarterly frequency. 09:40.930 --> 09:44.460 And again our starting point this the 14th of July. 09:45.070 --> 09:48.850 And we want to have a quarterly frequency for six periods. 09:50.800 --> 09:58.300 And if we get us a very first time step the Saudis offer September and typically a quarter is defined 09:58.330 --> 10:01.060 as the first quarter from January to March. 10:01.420 --> 10:08.500 Then the second quarter from April to June then the third quarter from July to September and the fourth 10:08.500 --> 10:11.110 quarter from October to December. 10:11.500 --> 10:17.500 And by default if you pass me a cure then we get to the end date of the quarter. 10:17.500 --> 10:26.000 So the next end date of a quarter is here starting from the 14th of July the 30th of September. 10:26.350 --> 10:30.580 So this is the very last day of the third quarter. 10:30.580 --> 10:38.140 And then the next time stems that consequently the thirty first of December to 15 and then the thirty 10:38.140 --> 10:44.390 first of March to 16 and so on and we can also see this year and the frequency. 10:44.650 --> 10:48.080 So if we pass cure then by default we are getting here. 10:48.090 --> 10:57.040 Q December so that means that the last month of the fourth quarter is December and bypassing here. 10:57.040 --> 11:03.200 Q S instead of Q We are getting actually the very first day off a quarter. 11:03.250 --> 11:11.820 So let's do this here and starting from the 14th of July the very first day of the next quarter is in 11:11.850 --> 11:13.670 the first of October. 11:13.740 --> 11:18.030 Then we get the first of January to 16 and so on. 11:18.030 --> 11:25.920 And in this case the frequency is called Q S January and now let's assume that we want to have quarters 11:26.010 --> 11:30.900 different from January to March April to June and so on. 11:30.900 --> 11:39.480 So let's assume for example that we want to have quarters from May to August and from September to November 11:40.050 --> 11:42.530 and we can do this by passing your flight sample. 11:42.540 --> 11:49.260 Q May and by default we are getting at the very last day of each quarter. 11:49.590 --> 11:56.700 So starting from the 14th of July the next last day of the quarter is the thirty first of August and 11:56.700 --> 11:57.310 so on. 11:57.900 --> 12:02.550 And also instead of getting the last day of each quarter we can get to the first day of each quarter 12:02.550 --> 12:10.860 by passing q s main and we should get then the first of August s first time. 12:11.420 --> 12:13.950 And this is true. 12:14.060 --> 12:22.570 All right now let's go to any other frequency and we are starting here also with the 14th of July. 12:22.670 --> 12:29.600 We want to have six periods and a annual frequency so we can pass here a while we can also pass your 12:29.600 --> 12:30.070 base. 12:30.080 --> 12:31.640 So this is the same 12:34.340 --> 12:39.140 and by default we are getting near the very last day of the calendar year. 12:39.200 --> 12:45.370 So here are the thirty first of December so we are getting here the thirty first of December off to 12:45.370 --> 12:49.150 15 16 17 18 19 and 20. 12:49.360 --> 12:57.340 And the frequency is called annually December and alternative we can also get the very first day of 12:57.370 --> 12:58.350 a calendar year. 12:58.360 --> 13:01.150 So the 1st of January. 13:01.420 --> 13:03.910 And therefore we have to pass a s 13:06.940 --> 13:09.130 and starting from the 14th of July. 13:09.160 --> 13:15.230 So the very first time stem is the first of January to a 16. 13:15.370 --> 13:18.440 And here we can see the frequency as a s. 13:18.500 --> 13:26.060 January All right now let's assume that we want to have annual data but instead of having a year from 13:26.060 --> 13:31.420 January to December we want to f a year from July to June. 13:32.000 --> 13:35.420 And therefore we can pass here a June 13:39.270 --> 13:41.070 A.B. get by default. 13:41.100 --> 13:47.420 The very last day of our year and the very last day is here the 30th of June. 13:47.880 --> 13:53.830 And starting from the 14th of July to 15 our first time stamp is now in 216. 13:53.880 --> 14:02.220 And of course we can also get the very first day of each year by passing year A S June. 14:02.300 --> 14:08.510 And if you want to have the year from July 2 to 1 with the very first day of the year then we have to 14:08.510 --> 14:09.920 pass you a s. 14:09.950 --> 14:10.570 July. 14:11.210 --> 14:13.550 So this should work. 14:13.550 --> 14:16.330 So now we have the 1st of July. 14:16.670 --> 14:24.410 And last but not least we can also create a yearly timestamps for example for the 24th of November. 14:24.440 --> 14:26.010 So this is my birthday. 14:26.270 --> 14:31.610 And let's assume I want to f the time stamps for the last 10 birthdays. 14:31.610 --> 14:37.640 So the last birthday was on the 24th of November to 18 and periods is 10. 14:37.970 --> 14:44.550 And here again I can pass the special method BD Date date offset to the frequency prior meter. 14:44.960 --> 14:50.770 And you're within the brackets like I can pass the frequency of one year. 14:50.930 --> 14:57.990 So let's try this out until we get to the day time index with the last 10 best days. 14:58.130 --> 15:07.310 So starting from the 24th of November 2009 to the 24th of November 2018. 15:07.360 --> 15:07.820 All right. 15:07.820 --> 15:14.270 This was a very detailed look on the date range method and are actually plenty of combinations that 15:14.270 --> 15:15.170 we can use here. 15:15.170 --> 15:19.530 So the date range method is quite powerful and useful. 15:19.880 --> 15:25.640 And in practice there's hardly any case that we cannot come up with the date range method. 15:25.640 --> 15:31.460 So this is very useful and with this we are finished with this video and I hope to see you also in the 15:31.460 --> 15:32.570 next one by.