CODEWITHSUNDEEP
rjoinconditional-statementsdplyr
chrischi
chrischi

Reputation: 185

Conditional Join with DPLYR

I am attempting to join two data frames by a variable that has a different name and by a single conditional criterion (i.e. the year) in the source.

Investigating various solutions has not provided me with an answer.

Data:

x_6.0_2011 <- data.frame(t = c("1", "2", "3", "7"), value_1 = c("0.9", "0.6", "0.1", "0.8"))

x <- data.frame(year = c("2010", "2011", "2012", "2013"), t.nr = c("1", "1", "2", "7"), value = c("0.2", "0.5", "0.7", "0.3"))

What I tried:

x2011 <-   ifelse(x$year == '2011', 
                  left_join(x, x_6.0_2011, by = c('t.nr' = 't')), 0)

--> producing a list - so wrong solution path

x20111 <- left_join(x, x_6.0_2011, by = c('t.nr' = 't'), ,year== 2011 )

--> producing a df but only replacing the correct value and coyping in the remainders from the x_6.0_2011 - so wrong solution path

xx_6.02011 <- left_join(filter(x, year == '2011'), x_6.0_2011, by = c('t.nr' = 't'))

--> producing df with one line but I would like to have kept the other entries not containing any value

Result Expected:

Yr       t.nr   value  value_1
2010     1       0.2    0
2011     1       0.3    0.9
2012     2       0.7    0
2013     7       0.3    0

... or similar at least in the last column not a value I cannot deduces.

The minimal example works however my working project (217 variable) still produces an error:

I do consistently get: Error: data_frames can only contain 1d atomic vectors and lists

Any advice is more than appreciated. Or are my attempts to simplistic?

Upvotes: 2

Views: 3245

Answers (1)

akrun
akrun

Reputation: 887731

You could use the devel version of data.table

library(data.table)#v1.9.5+
setDT(df1)[df2, on=c('t'='t.nr')][year!=2011, value_1:='0'][]
#    t value_1 year value
#1: 1       0 2010   0.2
#2: 1     0.9 2011   0.5
#3: 2       0 2012   0.7
#4: 7       0 2013   0.3

Upvotes: 4

Related Questions