create an OHLC series from ticker data using R

Question

This seems like it should be a common thing, but all my searching comes up with half or unfinished answers.

I have a set of data in a csv. But the data is set up so

Answer 1

from http://r.789695.n4.nabble.com/Convert-tick-data-to-OHLC-td926206.html

ohlc <- function(ttime,tprice,tvolume,fmt) {
    ttime.int <- format(ttime,fmt)
    data.frame(time = ttime[tapply(1:length(ttime),ttime.int,function(x) {head(x,1)})],
    .Open = tapply(tprice,ttime.int,function(x) {head(x,1)}), 
    .High = tapply(tprice,ttime.int,max),
    .Low = tapply(tprice,ttime.int,min),
    .Close = tapply(tprice,ttime.int,function(x) {tail(x,1)}),
    .Volume = tapply(tvolume,ttime.int,function(x) {sum(x)}),
    .Adjusted = tapply(tprice,ttime.int,function(x) {tail(x,1)}))
} 

Answer 2

Azoo series can be reclassed as OHLC with class(data) <- c('zoo', 'OHLC', 'some other class').

From ?quantmod::quantmod.OHLC:

‘quantmod.OHLC’ is actually just a renaming of an object of class ‘zoo’, with the convention of NAME.Open, NAME.High, ... for the column names.

If you don't have the volume column, then there is no code you can write that will add that data. It seems like you just have a series of intraday prices. You can still class these as zoo or xts and do time series analysis. If you want to reduce the resolution of your data, you could always scan over each day and use the functions first, max, min, last to get your O, H, L, C.

Answer 3

To illustrate @JoshuaUlrich's comment:

x <- read.csv(header=TRUE, stringsAsFactor=FALSE, 
              text="time,price,volume,
                    7/18/10 0:09,0.04951,20,
                    7/18/10 4:43,0.05941,50.01,
                    7/18/10 18:48,0.0808,5,
                    7/18/10 22:44,0.08585,10,
                    7/18/10 23:00,0.08584,5,
                    7/18/10 23:00,0.08584,5,
                    7/19/10 4:53,0.0909,5,
                    7/19/10 17:24,0.09307,80,
                    7/19/10 18:03,0.08911,100,
                    7/19/10 18:07,0.08752,100,")
# create an xts object
myxts <- xts(x[, 2:3], order.by=as.POSIXct(x[, 1], format='%m/%d/%y %H:%M'))
myxts

myxts
#                       price volume
# 2010-07-18 00:09:00 0.04951  20.00
# 2010-07-18 04:43:00 0.05941  50.01
# 2010-07-18 18:48:00 0.08080   5.00
# 2010-07-18 22:44:00 0.08585  10.00
# 2010-07-18 23:00:00 0.08584   5.00
# 2010-07-18 23:00:00 0.08584   5.00
# 2010-07-19 04:53:00 0.09090   5.00
# 2010-07-19 17:24:00 0.09307  80.00
# 2010-07-19 18:03:00 0.08911 100.00
# 2010-07-19 18:07:00 0.08752 100.00
to.minutes(myxts)
#                     myxts.Open myxts.High myxts.Low myxts.Close myxts.Volume
# 2010-07-18 00:09:00    0.04951    0.04951   0.04951     0.04951        20.00
# 2010-07-18 04:43:00    0.05941    0.05941   0.05941     0.05941        50.01
# 2010-07-18 18:48:00    0.08080    0.08080   0.08080     0.08080         5.00
# 2010-07-18 22:44:00    0.08585    0.08585   0.08585     0.08585        10.00
# 2010-07-18 23:00:00    0.08584    0.08584   0.08584     0.08584        10.00
# 2010-07-19 04:53:00    0.09090    0.09090   0.09090     0.09090         5.00
# 2010-07-19 17:24:00    0.09307    0.09307   0.09307     0.09307        80.00
# 2010-07-19 18:03:00    0.08911    0.08911   0.08911     0.08911       100.00
# 2010-07-19 18:07:00    0.08752    0.08752   0.08752     0.08752       100.00

to.minutes is one of many wrappers for to.period. This is equivalent, but more general:

to.period(myxts, 'minutes', 1)