CODEWITHSUNDEEP
rsubsetxts
Florent
Florent

Reputation: 1928

out-of-range error when subseting values from an xts object in R

I would like to subset dates from an xts object based on logical values in another xts object, but R returns an out-of-range error despite having in-range values.

For example I would like to filter dates and prices where RSI is above 60.

> strength <- RSI(d, 14)>60
> strength["2016-10-17::"]
             RSI
2016-10-17  TRUE
2016-10-18  TRUE
2016-10-19  TRUE
2016-10-20 FALSE
2016-10-21 FALSE

> d["2016-10-17::"]
               Open
2016-10-17 642.2760
2016-10-18 640.5988
2016-10-19 637.0000
2016-10-20 631.9800
2016-10-21 633.6470

> d["2016-10-17::"][strength == TRUE]
Error in `[.xts`(d["2016-10-17::"], strength == TRUE) : 
  'i' or 'j' out of range

This is not the output I expect because both my objects have data until 2016-10-21. What could be wrong? I would like something like :

> d["2016-10-17::"][strength == TRUE]
               Open
2016-10-17 642.2760
2016-10-18 640.5988
2016-10-19 637.0000

This is the str of my xts objects : 

> str(d)
An ‘xts’ object on 2013-09-02/2016-10-21 containing:
  Data: num [1:1146, 1] 127 128 121 121 116 ...
 - attr(*, "dimnames")=List of 2
  ..$ : NULL
  ..$ : chr "Open"
  Indexed by objects of class: [Date] TZ: UTC
  xts Attributes:  
List of 2
 $ dateFormat: chr "Date"
 $ na.action :Class 'omit'  atomic [1:92] 1 2 3 4 5 6 7 8 9 10 ...
  .. ..- attr(*, "index")= num [1:92] 1.37e+09 1.37e+09 1.37e+09 1.37e+09 1.37e+09 ...

> str(strength)
An ‘xts’ object on 2013-09-16/2016-10-21 containing:
  Data: logi [1:1132, 1] FALSE FALSE FALSE FALSE FALSE FALSE ...
 - attr(*, "dimnames")=List of 2
  ..$ : NULL
  ..$ : chr "RSI"
  Indexed by objects of class: [Date] TZ: UTC
  xts Attributes:  
 NULL
>

Thank you

Upvotes: 0

Views: 646

Answers (1)

FXQuantTrader
FXQuantTrader

Reputation: 6891

You didn't make a reproducible example, so here is some toy data. Your issue is that you did not subset strength with the same time window (so your inner strength == TRUE logical series has different row length to your d row length, generating your error. i.e. NROW(strength == TRUE) >> NROW(d["2016-10-17::"]) ):

library(quantmod)
getSymbols("AAPL")
d <- AAPL
strength <- RSI(Cl(d)) > 60

You shouldn't get an error if you do this:

d["2016-10-17::"][strength["2016-10-17::"] == TRUE]

Upvotes: 2

Related Questions