基于某些转换标准的数据转换

Question

I want to transform a dataset based on certain conditions. These conditions are given in another dataset. Let me explain it using an example.

Suppose I've a dataset in the following format:

Date       Var1 Var2
3/1/2016    8   14
3/2/2016    7   8
3/3/2016    7   6
3/4/2016    10  8
3/5/2016    5   10
3/6/2016    9   15
3/7/2016    2   5
3/8/2016    6   14
3/9/2016    8   15
3/10/2016   8   8

And the following dataset has the transformation conditions and is in the following format:

Variable    Trans1  Trans2
  Var1       1||2   0.5||0.7
  Var2       1||2   0.3||0.8

Now, I want to extract first conditions from transformation table for Var1, 1.0.5, and add 1 to Var1 and multiply it by 0.5. I'll do the same for var2, add by 1 and multiply by 0.3. This transformation will give me new variable Var1_1 and var2_1. I'll do the same thing for the other transformation, which will give me Var1_2 and Var2_2. For Var1_2, the transformation is Var1 sum with 2 and multiplied by 0.7.

After the transformation, the dataset will look like the following:

  Date     Var1 Var2    Var1_1  Var2_1  Var1_2  Var2_2
3/1/2016    8   14       4.5     4.5      7      11.2
3/2/2016    7   8        4       2.7      6.3     7
3/3/2016    7   6        4       2.1      6.3     5.6
3/4/2016    10  8        5.5     2.7      8.4     7
3/5/2016    5   10       3       3.3      4.9     8.4
3/6/2016    9   15       5       4.8      7.7    11.9
3/7/2016    2   5        1.5     1.8      2.8     4.9
3/8/2016    6   14       3.5     4.5      5.6    11.2
3/9/2016    8   15       4.5     4.8      7      11.9
3/10/2016   8   8        4.5     2.7      7       7

Answer 1

Given that your original data.frame is called df and your conditions table cond1 then we can create a custom function,

funV1Cond1 <- function(x){
  t1 <- as.numeric(gsub("[||].*", "", cond1$Trans1[cond1$Variable == "Var1"]))
  t2 <- as.numeric(gsub("[||].*", "", cond1$Trans2[cond1$Variable == "Var1"]))
  result <- (x$Var1 + t1)*t2
  return(result)
}
funV1Cond1(df)
 #[1] 4.5 4.0 4.0 5.5 3.0 5.0 1.5 3.5 4.5 4.5

Same way with function 2

funV1Cond2 <- function(x){
  t1 <- as.numeric(gsub(".*[||]", "", cond1$Trans1[cond1$Variable == "Var1"]))
  t2 <- as.numeric(gsub(".*[||]", "", cond1$Trans2[cond1$Variable == "Var1"]))
  result <- (x$Var1 + t1)*t2
  return(result)
}
funV1Cond2(df)
 #[1] 7.0 6.3 6.3 8.4 4.9 7.7 2.8 5.6 7.0 7.0

Assuming that Trans1 column has 3 conditions ie 1, 2, 3 then,

as.numeric(sapply(str_split(cond1$Trans1[cond1$Variable == "Var1"], ','),function(x) x[2]))
#[1] 2
as.numeric(sapply(str_split(cond1$Trans1[cond1$Variable == "Var1"], ','),function(x) x[1]))
#[1] 1
as.numeric(sapply(str_split(cond1$Trans1[cond1$Variable == "Var1"], ','),function(x) x[3]))
#[1] 3

Note that I changed the delimeter to a ','