Why all values become 1 after dcast?

Question

I am reshaping my dataframe as follows.

I want to set the Wavelength column as the column names with the code as follows.

dcast(unique(df1, by=cols), ID + Condition ~ Wavelength, value.var = 'mean')

But all the values became 1s as follows.

I would appreciate any help.

Yanpeng

Answer 1

That comment is right, it's a general programming question, and text would be much easier for people to help with rather than images.

But, since I ran into this not long ago...

Take a look at the message shown in blue just above the output table. My guess is you have some repeated ID/Condition/Wavelength combinations and it's defaulting to reporting the number of matching rows in the long-format data frame. You'll need to either make sure you have all unique combinations, or explicitly define what you expect it to do with multiple values going into a single output row/column position (like, median/mean/max/min/whatever). The unique() call won't do it since varying mean values would still make for unique input rows.

For example compare this:

> reshape2::dcast(data.frame(
+     Wavelength=c(664, 666, 662),
+     Condition="late",
+     ID=c(6221, 6222, 6223),
+     mean=c(106817, 110357, 112998)),
+   ID + Condition ~ Wavelength, value.var = "mean")
    ID Condition    662    664    666
1 6221      late     NA 106817     NA
2 6222      late     NA     NA 110357
3 6223      late 112998     NA     NA

with this:

> reshape2::dcast(data.frame(
+     Wavelength=c(664, 664, 666, 662),
+     Condition="late",
+     ID=c(6221, 6221, 6222, 6223),
+     mean=c(106800, 106817, 110357, 112998)),
+   ID + Condition ~ Wavelength, value.var = "mean")
Aggregation function missing: defaulting to length
    ID Condition 662 664 666
1 6221      late   0   2   0
2 6222      late   0   0   1
3 6223      late   1   0   0