Obtenez la valeur la plus élevée dans la plage horaire tous les jours (xts)

Question

L'objet XTS XTS1 a environ trois jours de données par intervalles de 15 minutes. Je veux trouver les valeurs maximales enregistrées tous les jours entre 900 heures et 9h30, puis enregistrer ces valeurs sur un vecteur. Comme indiqué dans l'exemple.

Je ferais alors la même chose pour le minimum. L'objectif est d'enregistrer la fourchette d'ouverture quotidienne d'un instrument financier (intervalle d'ouverture journalier = maximum quotidien & min pendant une période donnée). Pour rester simple, j'ignore le minimum sur cet exemple.

La question est très similaire sinon une copie de this one. Pourtant, la solution fournie ici est de trouver des maximums hebdomadaires dans une base de données avec fréquence journalière, alors que celle-ci sert à trouver des maximums horaires dans une base de données avec fréquence intrajournalière. Plus important encore, la solution fournie ne fonctionne pas dans ce cas (voir ci-dessous).

library('xts') 
XTS1 <- structure(c(0.2284, 0.2283, 0.2284, 0.2288, 0.2307, 0.2319, 0.2322, 0.2327, 0.2328, 0.2327, 0.2327, 0.2327, 0.2331, 0.2338, 0.2338, 0.2346, 0.2348, 0.2348, 0.236, 0.2361, 0.2362, 0.236, 0.2365, 0.2364, 0.2363, 0.236, 0.2359, 0.2364, 0.2365, 0.2363, 0.2363, 0.2362, 0.2363, 0.2364, 0.2365, 0.2365, 0.2364, 0.2364, 0.2363, 0.2363, 0.2364, 0.2363, 0.2363, 0.2363, 0.2364, 0.2364, 0.2364, 0.2364, 0.2365, 0.2365, 0.2365, 0.2365, 0.2365, 0.2366, 0.2366, 0.2365, 0.2365, 0.2364, 0.2364, 0.2364, 0.2364, 0.2363, 0.2365, 0.2365, 0.2366, 0.2365, 0.2365, 0.2364, 0.2364, 0.2368, 0.2373, 
        0.2378, 0.2379, 0.2382, 0.2383, 0.2384, 0.2384, 0.2384, 0.2383, 0.2385, 0.2385, 0.2387, 0.2385, 0.2385, 0.2385, 0.2386, 0.2387, 0.2387, 0.24, 0.2394, 0.2396, 0.2394, 0.2398, 0.2394, 0.2396, 0.24, 0.24, 0.2401, 0.2397, 0.2399, 0.2492, 0.2398, 0.2203, 0.2104, 0.2104, 0.2404, 0.2398, 0.24, 0.2401, 0.2402, 0.2399, 0.24, 0.24, 0.2403, 0.2406, 0.2401, 0.2404, 0.2406, 0.2406, 0.2405, 0.241, 0.2411, 0.2411, 0.2411, 0.2411, 0.2411, 0.2411, 0.2411, 0.2411, 0.2411, 0.2411, 0.2411, 0.2411, 0.2411, 0.241, 0.2405, 0.2406, 0.2405, 0.2404, 0.2404, 0.2406, 0.2408, 0.2406, 0.2406, 0.2407, 0.2407, 0.2407, 0.2408, 0.2408, 0.2408, 0.2408, 0.2408, 0.2408, 0.2408, 0.2411, 0.241, 0.241, 0.2411, 0.2411, 0.2412, 0.2412, 0.2411, 0.2411, 0.2411, 0.2412, 0.2412, 0.2412, 0.2412, 0.2412, 0.2412, 0.2412, 0.2411, 0.2412, 0.2412, 0.2412, 0.2413, 0.2412, 0.2411, 0.2411, 0.241, 0.241, 0.2411, 0.2409, 0.2417, 0.2422, 0.2421, 0.2428, 0.2426, 0.2428, 0.2421, 0.2421, 0.2419, 0.2422, 0.2424, 0.2427, 0.2437, 0.2438, 0.2441, 0.2442), .Dim = c(199L, 1L), .indexCLASS = c("POSIXct", "POSIXt"), .indexTZ = "", tclass = c("POSIXct", "POSIXt"), tzone = "", class = c("xts", "zoo"), .CLASS = structure("double", class = "CLASS"), formattable = structure(list(formatter = "formatC", format = structure(list(format = "f", digits = 2), .Names = c("format", "digits")), preproc = "percent_preproc", postproc = "percent_postproc"), .Names = c("formatter", "format", "preproc", "postproc")), index = structure(c(1413981900, 1413982800, 1413983700, 1413984600, 1413985500, 1413986400, 1413987300, 1413988200, 1413989100, 1413990000, 1413990900, 1413991800, 1413992700, 1413993600, 1413994500, 1413995400, 1413996300, 1413997200, 1413998100, 1413999000, 1413999900, 1414000800, 1414001700, 1414002600, 1414003500, 1414004400, 1414005300, 1414006200, 1414007100, 1414008000, 1414009800, 1414010700, 1414011600, 1414015200, 1414016100, 1414017000, 1414017900, 1414018800, 1414019700, 1414020600, 1414021500, 1414022400, 1414023300, 1414024200, 1414025100, 1414026000, 1414026900, 1414027800, 1414028700, 1414029600, 1414030500, 1414031400, 1414032300, 1414033200, 1414034100, 1414035000, 1414035900, 1414036800, 1414037700, 1414038600, 1414039500, 1414040400, 1414041300, 1414042200, 1414043100, 1414044000, 1414044900, 1414045800, 1414046700, 1414047600, 1414048500, 1414049400, 1414050300, 1414051200, 1414052100, 1414053000, 1414053900, 1414054800, 1414055700, 1414056600, 1414057500, 1414058400, 1414059300, 1414060200, 1414061100, 1414062000, 1414062900, 1414063800, 1414064700, 1414065600, 1414066500, 1414067400, 1414068300, 1414069200, 1414070100, 1414071000, 1414071900, 1414072800, 1414073700, 1414074600, 1414075500, 1414076400, 1414077300, 1414078200, 1414079100, 1414080000, 1414080900, 1414081800, 1414082700, 1414083600, 1414084500, 1414085400, 1414086300, 1414087200, 1414088100, 1414089000, 1414089900, 1414090800, 1414091700, 1414092600, 1414093500, 1414094400, 1414096200, 1414097100, 1414098000, 1414101600, 1414102500, 1414103400, 1414104300, 1414105200, 1414106100, 1414107000, 1414107900, 1414108800, 1414109700, 1414110600, 1414111500, 1414112400, 1414113300, 1414114200, 1414115100, 1414116000, 1414116900, 1414117800, 1414118700, 1414119600, 1414120500, 1414121400, 1414122300, 1414123200, 1414124100, 1414125000, 1414125900, 1414126800, 1414127700, 1414128600, 1414129500, 1414130400, 1414131300, 1414132200, 1414133100, 1414134000, 1414134900, 1414135800, 1414136700, 1414137600, 1414138500, 1414139400, 1414140300, 1414141200, 1414142100, 1414143000, 1414143900, 1414144800, 1414145700, 1414146600, 1414147500, 1414148400, 1414149300, 1414150200, 1414151100, 1414152000, 1414152900, 1414153800, 1414154700, 1414155600, 1414156500, 1414157400, 1414158300, 1414159200, 1414160100, 1414161000, 1414161900, 1414162800, 1414163700, 1414164600, 1414165500, 1414166400, 1414167300), tzone = "", tclass = c("POSIXct", "POSIXt"))) 

#DESIRED OUTPUT 
         [,1] Max900.930am 
2014-10-22 08:45:00 0.2284   NA 
2014-10-22 09:00:00 0.2283  0.2283  
2014-10-22 09:15:00 0.2284  0.2284 
2014-10-22 09:30:00 0.2288  0.2288 
2014-10-22 09:45:00 0.2307  0.2288 
2014-10-22 10:00:00 0.2319  0.2288 
..... 

#This records max on desired time range, but in whole database rather than for every day 
max(XTS1['T09:00:00/T09:30:00']) 

#This says "Error: width > 0 is not TRUE" 
xxx = rollapply(data = XTS1, width = XTS1['T09:00:00/T09:30:00'], FUN = max) 

#This is derived from solution by @JoshuaUlrich to the other question 
#Doesn't work: records max on desired time range, in whole database rather than every day 
Max900.930am <- apply.daily(XTS1["T09:00/T09:30"], function(x) max(XTS1)) 
y <- merge(XTS1, Max900.930am, fill = na.locf) 

#Printout of Max900.930am, showing the problem 
         [,1] 
2014-10-22 09:30:00 0.2492 
2014-10-23 09:30:00 0.2492 
2014-10-24 09:30:00 0.2492 

Answer 1

Vous étiez très proche avec cette ligne de code:

Max900.930am <- apply.daily(XTS1["T09:00/T09:30"], function(x) max(XTS1)) 

Mais il y a une faute de frappe cruciale: max(XTS1) devrait être max(x). Cette erreur est la raison pour laquelle vous obtenez le maximum pour toute la série, pas seulement le jour. Cela aurait travaillé:

Max900.930am <- apply.daily(XTS1["T09:00/T09:30"], max) 

Et vous pouvez facilement obtenir à la fois au maximum et le minimum avec un appel:

range0900.0930 <- apply.daily(XTS1["T09:00/T09:30"], range) 

Answer 2

Solution:

library('data.table') 
Max900.930am <- apply.daily(XTS1, function(x) max(x['T09:00:00/T09:15:00'])) 
y <- merge(XTS1, Max900.930am, fill = na.locf) 
y$Max900.930am <- shift(y$Max900.930am, n=53, type=c("lead")) 

Merci à @jbaums pour cela.