使用列值的新数据帧中的子数据帧

Question

I am trying to subset a big data.frame in different data.frames automatically. I have something like:

Type1 Type2 Type3 Info1 Info2 Info3
A     1     Z     a     a     a    
A     2     Y     b     b     b
B     4     X     c     c     c
A     1     Z     d     d     d

I want to create a different data frame for each Type3>Type2>Type1

1st data frame (A1Z):

Type1 Type2 Type3 Info1 Info2 Info3
A     1     Z     a     a     a    
A     1     Z     d     d     d

2nd data frame (A2Y):

Type1 Type2 Type3 Info1 Info2 Info3
A     2     Y     b     b     b    

...

What is the fastest way to reach that? Thank you!!

Answer 1

You can use split to produce a list of data.frames:

> split(df, do.call(paste0, df[,1:3]))
$A1Z
  Type1 Type2 Type3 Info1 Info2 Info3
1     A     1     Z     a     a     a
4     A     1     Z     d     d     d

$A2Y
  Type1 Type2 Type3 Info1 Info2 Info3
2     A     2     Y     b     b     b

$B4X
  Type1 Type2 Type3 Info1 Info2 Info3
3     B     4     X     c     c     c

It's generally recommended to keep related data structures (like data.frames) in a single object (like a list in this case). If you want to assign those data.frames to the global environment you can investigate ?list2env .

---

As noted by Frank in the comments, splitting like this might not be necessary (although we can't say for sure without more information). With dplyr you could use group_by to group the data by those TypeX columns and then perform some manipulations to each group. You would start like this:

library(dplyr)
df %>% 
  group_by(Type1, Type2, Type3) %>% 
#   ... more manipulations...

#Source: local data frame [4 x 6]
#Groups: Type1, Type2, Type3
#
#  Type1 Type2 Type3 Info1 Info2 Info3
#1     A     1     Z     a     a     a
#2     A     2     Y     b     b     b
#3     B     4     X     c     c     c
#4     A     1     Z     d     d     d