Question
Thanks for the pointer to na.locf (Darren), updated example and results below:
I have tick data, which I have rolled into daily data, in order to calc daily volatility. Now that I have created the daily volatility, I would like to merge the daily data with the tick data again. However, I suspect the merge remains "empty" due to the index differences of the daily and tick data.
How would one merge the daily data with tick data?
Example:
AGL.xts <- xts(AGL_Frame[,-1], order.by=AGL_Frame[,1])
AGL.xts
Close
2012-01-19 16:46:11 32376
2012-01-19 16:46:32 32377
2012-01-19 16:46:32 32376
2012-01-19 16:46:42 32376
2012-01-19 16:46:42 32376
2012-01-19 16:46:42 32376
2012-01-19 16:46:45 32376
2012-01-19 16:46:48 32351
2012-01-19 16:46:54 32351
2012-01-19 16:46:57 32351
2012-01-19 16:46:57 32351
2012-01-19 16:47:14 32351
2012-01-19 16:47:14 32351
2012-01-19 16:47:19 32350
2012-01-19 16:47:32 32349
2012-01-19 16:47:32 32349
my.sample1 <- to.daily(AGL.xts[,1],1,'daily')
my.sample1
daily.Open daily.High daily.Low daily.Close
2011-12-01 17:00:27 31000 31479 30685 31350
2011-12-05 17:00:28 31225 31700 31015 31645
2011-12-06 17:00:22 31290 31626 31126 31500
2011-12-07 17:00:12 31550 31840 31215 31366
2011-12-08 17:00:09 31350 31875 31200 31200
2011-12-12 17:00:25 31093 31245 30310 30310
2011-12-13 17:00:24 30333 30767 30100 30430
2011-12-14 17:00:12 30210 30500 29575 29700
2011-12-19 17:00:03 29900 30005 29633 29679
my.AGL.roc <- ROC(my.sample1[,4])
my.AGL.sd <- apply.rolling(my.AGL.roc, FUN="sd", width=5)*sqrt(252)
my.AGL.sd
calcs
2011-12-05 17:00:28 NA
2011-12-06 17:00:22 NA
2011-12-07 17:00:12 NA
2011-12-08 17:00:09 NA
2011-12-12 17:00:25 0.2195421
2011-12-13 17:00:24 0.1966806
2011-12-14 17:00:12 0.2240305
2011-12-19 17:00:03 0.2327860
2011-12-20 17:00:28 0.2878848
2011-12-21 17:00:18 0.2275700
2011-12-22 17:00:12 0.2462184
2011-12-28 17:00:00 0.1633643
2011-12-29 17:00:20 0.1800739
2012-01-03 17:00:25 0.4068977
2012-01-04 17:00:13 0.3699694
2012-01-05 17:00:04 0.4014607
2012-01-09 17:00:05 0.4049482
2012-01-10 17:00:17 0.3934479
2012-01-11 17:00:07 0.2391906
2012-01-12 17:00:01 0.2328756
2012-01-16 17:00:02 0.2165803
2012-01-17 17:00:22 0.1910748
2012-01-18 17:00:19 0.1347729
2012-01-19 17:00:09 0.1198476
merged <- merge(AGL.xts,my.AGL.sd)
merged <- na.locf(merged)
merged
Close Calcs
2012-01-12 12:03:49 31920 0.2391906
2012-01-12 12:03:52 31920 0.2391906
2012-01-12 12:03:54 31920 0.2391906
2012-01-12 12:03:56 31941 0.2391906
2012-01-12 12:04:19 31910 0.2391906
2012-01-12 12:04:21 31910 0.2391906
2012-01-12 12:04:22 31909 0.2391906
2012-01-12 12:04:22 31903 0.2391906
2012-01-12 12:04:22 31910 0.2391906
2012-01-12 12:04:23 31910 0.2391906
2012-01-12 12:04:28 31910 0.2391906
2012-01-12 12:04:28 31910 0.2391906
2012-01-12 12:04:32 31910 0.2391906
2012-01-12 12:04:32 31910 0.2391906
2012-01-12 12:04:33 31909 0.2391906
2012-01-12 12:04:33 31910 0.2391906
2012-01-12 12:04:33 31910 0.2391906
2012-01-12 12:04:33 31910 0.2391906
2012-01-12 12:04:33 31910 0.2391906
2012-01-12 12:04:38 31901 0.2391906
This achieves my goal of using a daily indicator (5-day vol in this case) and applying it to ticks for analysis purposes. Thanks for the advice.
Solution
Items 14.5 and 14.6 in R Cookbook demonstrate merging monthly inflation data with daily IBM data, using merge
(with all=T
or all=F
depending on purpose), na.locf
and zoo
with seq
to generate a full set of dates (to cover dates when one or the other symbol has no data).
I've used the same approach to create blank 1m bars for minutes where there were no ticks, so I think it will work for merging daily and tick data too.
OTHER TIPS
not sure where the function apply.rolling
comes from, but it looks like its a rolling standard deviation with a lag of 5?
Well, you've got that it looks like. There are no values for the first five rows in calcs due to the implementation details of apply.rolling
.
But i'd agree with Joshua... not sure exactly what you're trying to do here...