I have two lists such as ;
type1 0 0 0 0 0 2 5 5 5 1 1 3
type2 0 0 0 0 0 1 5 2 3 0 0 1
here is the dput format if it can helps:
list(type1 = c(0,0,0,0,0,2,5,5,5,1,1,3),
type2 = c(0,0,0,0,0,1,5,2,3,0,0,1))
And I would like to transform it in dataframe such as:
Nb type Value
1 type1 2
2 type1 1
3 type1 1
4 type1 0
5+ type1 3
1 type2 0
2 type2 1
3 type2 1
4 type2 0
5+ type2 10
where df$type==type1 corresponds of the number of Nb 1,2,3,4 or >=5
and where df$type==type corresponds of the number of sum of Nb 1,2,3,4 or >=5 of its correspondents in type1 list.
For example:
in type1 we see 3 numbers >=5, then I had the row:
Nb type Value
5+ type1 3
, among those 3, I add the sum if type2, which is 5+2+3 = 10, then I add:
Nb type Value
5+ type1 3
5+ typ2 10
Does someone have an idea?
An option is to get the table on the 'type1' element after converting to a factor with levels specified as 0 to 5, do a group by sum of 'type2' elements where the group is the 'type1' element, stack them into two column data.frame, and add the 'type' column as well and rbind the list elements
out <- do.call(rbind, Map(cbind, type = names(list1), lapply(setNames(list(table(factor(list1$type1, levels = 0:5)), tapply(list1$type2, factor(list1$type1, levels = 0:5), FUN = function(x) sum(x, na.rm = TRUE))), names(list1)), stack))
)
row.names(out) <- NULL
out$values[is.na(out$values)] <- 0
subset(out[c(3, 1, 2)], ind != 0)
-output
ind type values
2 1 type1 2
3 2 type1 1
4 3 type1 1
5 4 type1 0
6 5 type1 3
8 1 type2 0
9 2 type2 1
10 3 type2 1
11 4 type2 0
12 5 type2 10
Or using forcats
library(forcats)
library(dplyr)
grp <- factor(with(list1, fct_collapse(as.character(type1),
`>=5` = as.character(type1)[type1 >=5])), levels = c(0:4, ">=5"))
v1 <- table(grp)
v2 <- tapply(list1$type2, grp, FUN = sum)
bind_rows(list(type1 = stack(v1)[2:1], type2 = stack(v2)[2:1]), .id = 'type') %>%
filter(ind != '0')