Concatenating, Merging, and Joining Part 1

Quick reference

• 00:03 Okay, in this video I want to talk about concatenating, merging, and joining.
• 00:09 So different ways we can combine different data frames together.
• 00:12 So let's say we have two or three data frames.
• 00:14 We want to combine them all into one data frame.
• 00:16 How do we do that? So
• 00:16 there's several different methods based on how we want to do it, all right.
• 00:21 And in this video, we're going to look at concatenating.
• 00:23 In the next couple of videos after this, we'll look at merging and joining.
• 00:26 So concatenating, and let me just type this out, concatenate.
• 00:30 It's kind of a weird word that some people haven't seen.
• 00:33 It just means kind of smushing together, right?
• 00:35 You're just combining two things.
• 00:36 And with concatenating, we're not doing any logic,
• 00:40 we're not merging these things based on the data that's inside of them.
• 00:44 We're just kind of slapping them together, right?
• 00:47 We don't really care.
• 00:47 So let me just paste in a couple of data frames here.
• 00:51 So we have df1 and df2, and we create them
• 00:55 the same way we've created data frames in the past with our pd.Dataframe function.
• 00:59 And inside here, I'm just going to use a Python dictionary because it's easier.
• 01:03 And so we have columns Monday, Tuesday.
• 01:06 And Wednesday, and Monday, Tuesday, and Wednesday.
• 01:09 So each of these data frames has the same columns.
• 01:11 And for the index, I just put a quick numbered index 1, 2, 3, 4.
• 01:15 For this one, this one's 5, 6, 7, 8.
• 01:17 So they have the same columns, but they have different indices, right?
• 01:20 So inside of here, for Mondays column we have M1, M2, M3, and M4.
• 01:26 For Tuesdays, we have T1, T2, T3, and T4.
• 01:29 For Wednesday, we have W1, W2, W3, and W4.
• 01:33 So I hope you see what I'm doing here.
• 01:34 I'm just creating columns of data that have the name of it.
• 01:38 W for Wednesday, T for Tuesday, and M for Monday.
• 01:42 For the second one, we're going to have the same format.
• 01:45 The same M for Monday, T for Tuesday, and
• 01:47 W for Wednesday, but the numbers are going to be different.
• 01:49 So these were 1, 2, 3, and 4.
• 01:51 These will all be 5, 6, 7, and 8, right?
• 01:54 And, of course, different indexes.
• 01:57 So let's go ahead and just run these and see what we've got here.
• 02:02 So we've got df1 and df2, Shift + Enter to run.
• 02:06 And so we see Monday, Tuesday, Wednesday.
• 02:07 Monday, Tuesday, Wednesday, they both have the same columns,
• 02:09 different indexes, right?
• 02:10 1, 2, 3, 4, 5, 6, 7, 8.
• 02:13 And the data is slightly different here at M1, M2, M3, M4, and
• 02:18 then M5, M6, M7, and M8.
• 02:21 So okay, just very basic data frames here with very basic data.
• 02:25 And so we just want to combine these two into one data frame.
• 02:30 How do we do it? Well, it's actually very easy.
• 02:32 We can just use the PD.concat function.
• 02:35 And inside of here we can just pass in a list of whatever data frames we want to
• 02:39 concatenate.
• 02:40 So we want df1 and df2.
• 02:43 When we run this, we get one big data frame.
• 02:46 And you'll notice we're sort of doing this on the row axis, the axis of 0.
• 02:55 Remember when we created columns way back when and we had to designate which axis
• 03:00 to put it on, the top ones here that was axis 1, the rows were axis 0.
• 03:04 By default, the concat function, combines on axis 0, which are these rows.
• 03:11 So we have 1, 2, 3, 4, 5, 6, 7, 8.
• 03:13 And the data is just these two data frames smush together.
• 03:18 So that's pretty cool.
• 03:20 It kind of makes sense to do these all on the axis because we have 1,
• 03:25 2, 3, 4, 5, 6, 7, 8.
• 03:27 It just seems to go together.
• 03:28 But if you wanted to do it by column, you could, you could just come up here and
• 03:32 designate the different axis.
• 03:34 So you can go axis = 1.
• 03:36 Remember, our columns are axis 1.
• 03:38 And look what we have here.
• 03:39 It's kind of interesting.
• 03:40 It's combining these two data frames, right?
• 03:43 But since we're doing it on this top axis,
• 03:47 the axis 1, each of these data frames has no data.
• 03:52 Because our first data frame doesn't have anything in row 5, 6, 7 and 8, right?
• 03:58 So we have no values for those.
• 04:01 The same thing with our second data frame.
• 04:05 It doesn't have anything for rows 1,
• 04:07 2, 3, and 4, so we get a bunch of nulls here, right?
• 04:11 So kind of an interesting thing to do it.
• 04:14 If you want to designate 0, you can and we just get the same thing we got earlier.
• 04:19 But the axis of 0 is the one by default.
• 04:22 So, you don't have to put that in, you could just run it like this.
• 04:24 So, that's concatenating.
• 04:26 Pretty simple.
• 04:27 It's just think of smushing them together.
• 04:28 I know smush is not a real word, but that's what I think of.
• 04:32 You're just combining two things, smushing them together.
• 04:35 We're not merging or joining based on any logic or anything like that.
• 04:39 We're just slapping these two things together.
• 04:42 And we use two data frames, you could use three, you could use 20,
• 04:45 it doesn't matter.
• 04:45 Any more than one is all you need to concatenate.
• 04:49 So that's cool.
• 04:51 So that's concatenating.
• 04:52 In the next video, we'll look at merging and joining.

Lesson notes are only available for subscribers.