CODEWITHSUNDEEP
rzoo
user3519324
user3519324

Reputation: 355

Delete rows from zoo-object based on conditions

I merged 2 different (irregular) time series (merge.zoo() ). Looks like: 

                     quir   schme 
2006-01-01 01:15:00 725.24   4.64
2006-01-01 01:30:00 725.24   4.64
2006-01-01 01:45:00      0   4.64
2006-01-01 02:00:00 725.24   4.64

I want to delete all rows where 0 values appear, to receive:

                    quir   schme 
2006-01-01 01:15:00 725.24   4.64
2006-01-01 01:30:00 725.24   4.64
2006-01-01 02:00:00 725.24   4.64

I didn't find a proper way to to this for zoo-objects so I used the very unelegant way to export the data to a *.csv and treat the rows upon re-import. Anyone who is willing to share a more elegant way?

Best regards Jochen

Upvotes: 1

Views: 1528

Answers (1)

akrun
akrun

Reputation: 887048

You can use subset or [

library(zoo) 
subset(z1, quir!=0)
#                      quir schme
#2006-01-01 01:15:00 725.24  4.64
#2006-01-01 01:30:00 725.24  4.64
#2006-01-01 02:00:00 725.24  4.64

If you need to remove rows if there are 0 values in any of the columns

z1[!rowSums(z1==0),]
#                      quir schme
#2006-01-01 01:15:00 725.24  4.64
#2006-01-01 01:30:00 725.24  4.64
#2006-01-01 02:00:00 725.24  4.64

Or as @G.Grothendieck mentioned in the comments, we can loop over the rows with apply and get the logical index with all or any. Here, ! indicates negation. !! double negation (z1!=0 and !!z1 are the same).

 z1[apply(z1 != 0, 1, all), ]
 z1[apply(!!z1, 1, all), ]
 z1[!apply(z1 == 0, 1, any), ]

data

z1 <- structure(c(725.24, 725.24, 0, 725.24, 4.64, 4.64, 4.64, 4.64
), .Dim = c(4L, 2L), .Dimnames = list(NULL, c("quir", "schme"
)), index = structure(c(1136096100, 1136097000, 1136097900, 1136098800
), class = c("POSIXct", "POSIXt"), tzone = ""), class = "zoo")

Upvotes: 2

Related Questions