how to turn vectors to matrix in R with unequal length

Question

I have vector with following structure like this. Vector with name "data_1" are numbers between cells "data_1" and "data_2" , etc

row name   cell meaning
1           data_1
2           3.4
3           2.3
...         ...
40          data_2
41          12
...         ...
60          data_3
61          63.2
...         ...

I want to create array of data frames DF like this. So each vector is in separate data frame.

DF[1] 
3.4
2.3
...

DF[2]  
12
...

DF[3]  
63.2
...

What is the fast way to do it? I have 4 million records, so traditional cycles such as for(i in 1:nrow) are too slow. Thanks.

Answer 1

Did you mean to have something like this?

library(zoo)

#split dataframe in list of values
df_list <- split(df, na.locf(sapply(df$cell_meaning, function(x) ifelse(grepl("[0-9]+\\.?[0-9]+",x),
                                                             NA,
                                                             which(df$cell_meaning==x)))))
names(df_list) <- NULL

#remove data_*
final_list <- lapply(df_list, `[`, -1,)

#final output
final_list
final_list[1]

Output is:

> final_list
[[1]]
[1] "3.4" "2.3"

[[2]]
[1] "12"

[[3]]
[1] "63.2"


> final_list[1]
[[1]]
[1] "3.4" "2.3"

Sample data:

df <- structure(list(cell_meaning = c("data_1", "3.4", "2.3", "data_2", 
"12", "data_3", "63.2")), .Names = "cell_meaning", class = "data.frame", row.names = c(NA, 
-7L))