1 00:00:02,820 --> 00:00:11,040 If you ever need to extract a table from a PDF file and save the table in a proper tabular format, 2 00:00:11,040 --> 00:00:14,370 such as a CSB file, then this is how you do it. 3 00:00:15,630 --> 00:00:21,300 And this video from this table in this PDA of documents, you can find this document attached in the 4 00:00:21,300 --> 00:00:22,230 lecture resources. 5 00:00:22,470 --> 00:00:30,090 So I'll go to my programming environments and I've uploaded the documents, the PDA of documents here 6 00:00:30,510 --> 00:00:32,009 in my project directory. 7 00:00:32,340 --> 00:00:40,080 So this might mean that's profile, and I'm going to import the library that extracts tables from PDF 8 00:00:40,080 --> 00:00:40,500 files. 9 00:00:41,550 --> 00:00:45,090 Now you'll need to install this with PIP install. 10 00:00:46,750 --> 00:00:49,450 So I am in the shell here, not in the council. 11 00:00:50,230 --> 00:00:51,850 So you need to write time, you lot. 12 00:00:53,150 --> 00:01:01,130 That's why and then imports Tabula, so I showed you that because you shouldn't install tabula, but 13 00:01:01,130 --> 00:01:03,200 tabula dash P. Y. 14 00:01:04,010 --> 00:01:05,960 So that's the proper installation. 15 00:01:07,180 --> 00:01:07,960 Execute, that's. 16 00:01:09,430 --> 00:01:14,950 Was that library installed then we are ready to write the code, which is very minimalistic. 17 00:01:16,150 --> 00:01:25,300 All you have to do is create a variable such as table and then write tabular dot reads PDF. 18 00:01:25,840 --> 00:01:28,540 Now this method gets through arguments. 19 00:01:28,870 --> 00:01:37,060 The first one being the path to the PDF document, which is whether the PDF and the second argument 20 00:01:37,060 --> 00:01:44,410 is the number of the page where the table is located, so pages equal to one. 21 00:01:45,600 --> 00:01:52,110 That is the first page in the documents, if you had a table in the second page, then you want to also 22 00:01:52,110 --> 00:02:00,600 have a right to hear or if you have multiple pages, you have to write different several read PDF lines. 23 00:02:00,900 --> 00:02:05,280 So like that one and to all implement a loop. 24 00:02:05,940 --> 00:02:15,030 So for I in range one to 10, if you have 10 pages, for example, and then writes, I hear so. 25 00:02:16,160 --> 00:02:21,860 The standard way of looking both in this case, I just want to show you a minimalistic example. 26 00:02:22,280 --> 00:02:29,510 So page number one and then putting the table run. 27 00:02:30,800 --> 00:02:31,610 And there we go. 28 00:02:33,530 --> 00:02:35,060 Yeah, that's the the output. 29 00:02:36,710 --> 00:02:44,780 Year, month and day out temperature pressure for her noid, so we have these columns here and each 30 00:02:44,780 --> 00:02:50,420 of them have built up not what type of object is this table? 31 00:02:50,780 --> 00:02:59,060 Well, you see a listes here and opening brackets, which means that this is a list, but this list 32 00:02:59,060 --> 00:03:00,500 has only one object. 33 00:03:01,430 --> 00:03:03,650 That object is a data frame. 34 00:03:04,010 --> 00:03:07,370 So the output is a list of data frames. 35 00:03:07,610 --> 00:03:12,710 If we had more than one tables, we would have a list of more than one data frames. 36 00:03:12,950 --> 00:03:15,770 In this case, we have only one list with one item. 37 00:03:16,160 --> 00:03:18,560 That item is for single data frame. 38 00:03:20,180 --> 00:03:29,420 So that means if you just do here, table zero, you stroked the table without the list, so that's 39 00:03:29,420 --> 00:03:30,680 the second ultimate we got. 40 00:03:31,280 --> 00:03:32,810 You see, there's no list here. 41 00:03:33,050 --> 00:03:40,730 And if you do print type of that table table zero. 42 00:03:41,790 --> 00:03:45,270 You see that this is a pandas dataframe in here. 43 00:03:46,470 --> 00:03:53,820 So we can make use of that and convert this table to table zero. 44 00:03:54,360 --> 00:04:01,200 We can convert it to Sears V. So to Sears V. is a method of pandas. 45 00:04:01,380 --> 00:04:10,590 So since this is a pandas dataframe because tabula, this library installed pandas on the big grounds. 46 00:04:11,130 --> 00:04:14,010 So it is able to generate these tables. 47 00:04:14,250 --> 00:04:18,529 So since this is a data sharing table, it does have access to sources methods. 48 00:04:19,140 --> 00:04:21,360 So defined the outputs. 49 00:04:22,450 --> 00:04:28,470 File path out from the CSB, for example, and run. 50 00:04:30,070 --> 00:04:33,400 And you should expect a certain file being generated in here. 51 00:04:33,680 --> 00:04:38,440 No, we see that we do have this comma here. 52 00:04:39,520 --> 00:04:48,370 So Python, what Python does is it adds an index column on the Left of the Year column here. 53 00:04:49,470 --> 00:04:55,710 And you can see that in here, so it's added this zero one two three four five six. 54 00:04:55,920 --> 00:05:01,240 If you don't want that, then you can specify an index is equal to none. 55 00:05:01,260 --> 00:05:06,180 So you don't want an index column around that and that will override the previous output. 56 00:05:06,180 --> 00:05:06,960 That's useful. 57 00:05:06,960 --> 00:05:10,770 And now you don't have that index column, so. 58 00:05:13,180 --> 00:05:18,640 Those are the dots, as you see the loss of row here, you have emptied out Tacoma, Tacoma, comma, 59 00:05:18,650 --> 00:05:24,070 comma, so no data, but we have 5.0, which reflects the formats here. 60 00:05:24,070 --> 00:05:33,430 So we don't have data in this last row, but we do have this five in the last column and that is reflected 61 00:05:34,030 --> 00:05:36,190 quite well in the design here. 62 00:05:37,010 --> 00:05:38,470 Oh, that is a complete code. 63 00:05:38,710 --> 00:05:39,550 Thank you for calling. 64 00:05:39,580 --> 00:05:40,360 I'll talk to you later.