How can I merge three columns into one column according to the order of in the three columns

Question

Could you give me an advice, please?

I have three columns, e.x.:

x     y     z
0.2   0.5   0.26
0.75  0.58  0.25
0.78  0.86  0.95
0.56  0.75  0.52
0.45  0.47  0.57
0.2   0.58  0.98

Result: And I need only one column with this sequence:

    xyz
x1  0.2
y1  0.5
z1  0.26
x2  0.75
y2  0.58
z2  0.25
... 0.78
    0.86
    0.95
    …

I didn´t find a similar problem... Thank You very much. I tried to use "dcast" (R) or "concat" (Python) but I am stack.

`pandas` `pivot` is probably what you're after https://pandas.pydata.org/pandas-docs/stable/generated/pandas.DataFrame.pivot.html — Andrew, Dec 18 '18 at 08:56
Possible duplicate of [How to pivot a dataframe](https://stackoverflow.com/questions/47152691/how-to-pivot-a-dataframe) — Andrew, Dec 18 '18 at 08:56
@Andrew this is actually a `melt` problem, the opposite of `pivot`. — yatu, Dec 18 '18 at 09:34

yatu · Accepted Answer · 2018-12-18T09:30:39.837

You can use pd.melt and create the row names you want with np.tile:

m = df.T.melt(value_name='xyz')
rows = np.tile(df.columns, df.shape[0])
m['variable'] = rows.tolist() + (m.variable + 1).astype(str)

    variable   xyz
0        x1  0.20
1        y1  0.50
2        z1  0.26
3        x2  0.75
4        y2  0.58
5        z2  0.25
6        x3  0.78
7        y3  0.86
8        z3  0.95
9        x4  0.56
10       y4  0.75
11       z4  0.52
12       x5  0.45
13       y5  0.47
14       z5  0.57
15       x6  0.20
16       y6  0.58
17       z6  0.98

And if you want is as index do:

m.set_index('variable')

score 0 · Answer 2 · answered Dec 18 '18 at 08:59

0

as.vector(unlist(t(df)))

This should do the trick. If you want the result by columns get rid of the t() operator

edit

Sorry I have assumed your columns belong to a dataframe. If this is not the case, you can create one using

df=cbind(x,y,z)

answered Dec 18 '18 at 08:59

boski

2,437
1
14
30

Lester_wu · Answer 3 · 2018-12-18T14:10:12.803

0

if x, y, z are list object, you can use zip function to retrieve element from these lists simultaneously, and use enumerate function to get the how many index in these lists.

for i, (a, b, c) in enumerate(zip(x, y, z)):
    print("x{} {}".format(i, a))
    print("y{} {}".format(i, b))
    print("z{} {}".format(i, c))

edited Dec 18 '18 at 14:10

answered Dec 18 '18 at 09:00

Lester_wu

151
5

Add some explanation perhaps? – Vineeth Sai Dec 18 '18 at 10:39

LOrD_ARaGOrN · Answer 4 · 2018-12-18T09:18:41.377

0

First stack the columns and then drop the multiindex:

df2 = df.stack().reset_index(drop=True)

Now sort the values

print df2.sort_values().reset_index(drop=True)

After this you can name the column.

edited Dec 18 '18 at 09:18

answered Dec 18 '18 at 09:13

LOrD_ARaGOrN

3,884
3
27
49

score 0 · Answer 5 · edited Dec 18 '18 at 11:51

x <- c(0.2, 0.75, 0.78, 0.56, 0.45, 0.2)
y <- c(0.5, 0.58, 0.86, 0.75, 0.47, 0.58)
z <- c(0.26, 0.25, 0.95, 0.52, 0.57, 0.98)

d <- data.frame(x,y,z)


datalist <- list()
for( i in 1:nrow(d))
{
  m <- d[i,]
  datalist[[i]] <- m
}
output <- unlist(datalist,use.names = F)
output

output

0.20 0.50 0.26 0.75 0.58 0.25 0.78 0.86 0.95 0.56 0.75 0.52 0.45 0.47 0.57 0.20
0.58 0.98

Hope this will work for you.

Regards, Dinesh Vasu

How can I merge three columns into one column according to the order of in the three columns

5 Answers5