Displaying 2 results from an estimated 2 matches for "daily_dates".
2012 Mar 05
1
index instead of loop?
...xx1 = merge(dt,ua_rd,by.x=1,by.y= 'rt_date',all.x=T)
xx = as.character(xx1[,2])
values <- c(NA, xx[!is.na(xx)])
ind = cumsum(!is.na(xx)) + 1
y <- values[ind]
if(ck_rows == length(y)){
fin = data.frame(fin,y)
}else{
bad = c(bad,tkr)
}
}
colnames(fin) = c('daily_dates',nms)
print("over all time for loop")
print(Sys.time()-start_t_all)
print(fin)
Thanks,
Ben
PS - the real/over-all issue is below, but it is probably too involved to
follow.
On Sat, Mar 3, 2012 at 2:30 PM, Ben quant <ccquant@gmail.com> wrote:
> Hello,
>
> Thank y...
2012 Mar 03
0
removing data look-ahead, something faster.
...xx1 = merge(dt,ua_rd,by.x=1,by.y= 'rt_date',all.x=T)
xx = as.character(xx1[,2])
values <- c(NA, xx[!is.na(xx)])
ind = cumsum(!is.na(xx)) + 1
y <- values[ind]
if(ck_rows == length(y)){
fin = data.frame(fin,y)
}else{
bad = c(bad,tkr)
}
}
colnames(fin) = c('daily_dates',nms)
# after this I would slice and dice the data into weekly, monthly, etc.
periodicity as needed, but this leaves it in daily format which is as
granular as I will get.
print("over all time for loop")
print(Sys.time()-start_t_all)
Regards,
Ben
[[alternative HTML version delet...