Skip to content

Error in `[.xts`(one, trim:length(two), ) : subscript out of bounds

2 messages · Stephen Choularton, Joshua Ulrich

Original
Stephen Choularton
# 
Hi

I am getting this error even though I'm pretty sure I am not out of bounds:

Error in `[.xts`(one, trim:length(two), ) : subscript out of bounds
In addition: Warning message:
In trim:length(two) :
 ? numerical expression has 305 elements: only the first used

I have tried all sorts of things but clearly I am not figuring out what 
is going wrong.

I am developing code to recover prices from IB and test for cointegration.

There are two files attached:

* checkPairFromIB

* metals.csv

It all worked fine until I discovered gold and silver returned data of 
different lengths:? Gold had 320 days of data and silver 305.? The calls 
for data at lines 25 and 26 are identical so I don't know why.? Can 
anyone shed light on this?

I then tried to trim the longer table by disposing of the earlier 15 
lines (see lines 44 to 55 of the code).? This is where the error 
messages came up.? Can anyone point out what I am doing wrong?

Thanks.

Stephen
Joshua Ulrich
# 
On Tue, Jul 10, 2018 at 3:14 PM, Stephen Choularton
<stephen at organicfoodmarkets.com.au> wrote:
I see a few problems in your code.  Note that I cannot replicate your
error, but I assume 'one' and 'two' are xts objects with OHLCV data.

The calls to length() will return the total number of _elements_ in
the xts objects, not the number of rows (or observations).  You likely
want the latter.

You define 'trim' as 'one - two'.  That is, you've subtracted two xts
objects.  Therefore, 'trim' contains the price differences of 'one'
minus 'two'.  Your code seems to treat 'trim' as the difference in the
number of observations between 'one' and 'two'.

The calls to `[.xts` treat the first argument as the locations of the
rows you want returned.  You can probably see why this is an error.
Creating a sequence that starts at "an xts object of price differences
between two securities" and ends at "the number of elements in an xts
object" does not make sense.

Since you only need close prices for the call to 'ca.jo()', I suggest
you do an inner join of the close prices of the two xts objects.

closePrices <- merge(one$Close, two$Close, join = "inner")
jotest <- ca.jo(closePrices, ...)