Suppose there is a database where:
x y z
a 2015 122.4
a 2015 122.4
b 2016 200.5
a 2014 300.6
c 2016 80.1
What I was wondering is to do a programming in the R that transforms the repeated values of z of each group x and y into 0, keeping a value of z, the final result being:
x y z
a 2015 122.4
a 2015 0
b 2016 200.5
a 2014 300.6
c 2016 80.1
How to do this?
See the duplicated function.
dados$z[duplicated(dados$z)] <- 0
dados
x y z
1 a 2015 122.4
2 a 2015 0.0
3 b 2016 200.5
4 a 2014 300.6
5 c 2016 80.1
If you want to keep the original base, make a copy in advance and modify the copy, dados2 <- dados , followed by the above code (with dados2 ).
Data:
dados <-
structure(list(x = structure(c(1L, 1L, 2L, 1L, 3L), .Label = c("a",
"b", "c"), class = "factor"), y = c(2015L, 2015L, 2016L, 2014L,
2016L), z = c(122.4, 0, 200.5, 300.6, 80.1)), .Names = c("x",
"y", "z"), row.names = c(NA, -5L), class = "data.frame")
EDITION.
After looking at Rafael Cunha's answer I noticed that a fundamental point of the question had failed, that repeated or duplicate values should be grouped by columns x and y . The following answer addresses this PO requirement.
dados$z <- ave(dados$z, dados$x, dados$y, FUN = function(.z)
ifelse(duplicated(.z), 0, .z))
Following code, using the dplyr package. In this case, I take into account the grouping of x and y
library(dplyr)
dados <-
structure(list(
x = structure(c(1L, 1L, 2L, 1L, 3L),
.Label = c("a", "b", "c"), class = "factor"),
y = c(2015L, 2015L, 2016L, 2014L, 2016L),
z = c(122.4, 122.4, 200.5, 300.6, 80.1)),
.Names = c("x", "y", "z"), row.names = c(NA, -5L), class = "data.frame")
dados %>%
group_by(x, y) %>%
transmute(z = ifelse(duplicated(z), 0, z))