Fusionner plusieurs trames de données sur deux colonnes communes
https://stackoverflow.com/questions/4199798
italiano
english
français
española
中国
日本の
العربية
Deutsch
한국어
Português
RussianQuestion
Je l'ai vu quelques questions sur la fusion des fichiers csv dans une trame de données. Que faire si les trames de données sont déjà dans l'espace de travail. J'ai cinq grandes zoos que je CAST trames de données, puis fondre. Voici la tête d'un:
> head(df.mon.ssf.ret)
date variable value
1 2009.000 AA1C NA
2 2009.083 AA1C NA
3 2009.167 AA1C NA
4 2009.250 AA1C NA
5 2009.333 AA1C NA
6 2009.417 AA1C NA
Je pourrais fusionner ces derniers sur « date » et « variable » avec une série de fusions imbriquées, mais qui semble maladroite. Y at-il une manière plus programmatique de fusion?
Si je suis convaincu que les colonnes sont dans le même ordre dans tous les zoos, puis-je être sûr que l'état fondu soutient que cbind de commande et de l'utilisation? Merci!
Mise à jour:
Il y a quelque chose qui me manque sur la philosophie d'utilisation de la fonte. Voici ce qui se passe quand je fusionner comme un zoo et faire fondre comme trame de données très large en utilisant trois des zoos:
> temp <- merge(z.ssf.oi, z.ssf.oig, z.ssf.ret)
> class(temp)
[1] "zoo"
> temp2 <- cbind(index(temp), as.data.frame(temp))
> class(temp2)
[1] "data.frame"
> names(temp2)[1] <- "date"
> dim(temp2)
[1] 12 1204
> temp3 <- melt(temp2, id="date")
Error in data.frame(ids, variable, value) :
arguments imply differing number of rows: 12, 14436
> head(temp2)[, 1:5]
date AA1C.z.ssf.oi AAPL1C.z.ssf.oi ABT1C.z.ssf.oi ABX1C.z.ssf.oi
Jan 2009 Jan 2009 1895.800 49191.25 NA NA
Feb 2009 Feb 2009 1415.579 42650.26 NA 6267.96
Mar 2009 Mar 2009 1501.398 36712.20 NA 11581.65
Apr 2009 Apr 2009 1752.936 74376.27 NA 12168.29
May 2009 May 2009 1942.874 96307.30 NA 13490.60
Jun 2009 Jun 2009 NA 79170.70 NA 16337.21
Mise à jour 2: Merci pour l'aide! Voici une solution très manuelle
> A <- cbind(index(z.ssf.oi), as.data.frame(z.ssf.oi))
> names(A)[1] <- "date"
> B <- cbind(index(z.ssf.oig), as.data.frame(z.ssf.oig))
> names(B)[1] <- "date"
> C <- cbind(index(z.ssf.ret), as.data.frame(z.ssf.ret))
> names(C)[1] <- "date"
> A.melt <- melt(A, id="date")
> head(A.melt)
date variable value
1 Jan 2009 A1C NA
2 Feb 2009 A1C NA
3 Mar 2009 A1C NA
4 Apr 2009 A1C NA
5 May 2009 A1C NA
6 Jun 2009 A1C NA
> B.melt <- melt(B, id="date")
> C.melt <- melt(C, id="date")
> ans <- merge(merge(A.melt, B.melt, by=c("date", "variable")), C.melt, by=c("date", "variable"))
> names(ans)[3:5] <- c("oi", "oig", "ret")
> head(ans)
date variable oi oig ret
1 Apr 2009 A1C NA NA NA
2 Apr 2009 AA1C NA NA NA
3 Apr 2009 AAPL1C 59316.88 0.3375786 0.008600073
4 Apr 2009 ABB1C NA NA NA
5 Apr 2009 ABT1C NA NA NA
6 Apr 2009 ABX1C NA NA NA
(et les agences nationales sont d'un ensemble de données incomplètes à la maison et qui ont besoin du cadran dans le filtrage de ma base de données)
Mise à jour 3: Voici quelques dputs (je pris le sous-ensemble de chaque grand zoo [1:10, 1:10] et converties en trames de données)
> dput(A)
structure(list(group = structure(c(1L, 1L, 1L, 1L, 1L, 1L, 1L,
1L, 1L, 1L), class = "factor", .Label = "oi"), date = structure(c(2009,
2009.08333333333, 2009.16666666667, 2009.25, 2009.33333333333,
2009.41666666667, 2009.5, 2009.58333333333, 2009.66666666667,
2009.75), class = "yearmon"), AA1C = c(NA_real_, NA_real_, NA_real_,
NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_
), AAPL1C = c(49226.391, 42662.1589473684, 35354.4254545455,
57161.6495238095, 84362.895, NA, NA, 47011.8519047619, 57852.2171428571,
33058.0090909091), ABT1C = c(NA_real_, NA_real_, NA_real_, NA_real_,
NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_),
ABX1C = c(NA_real_, NA_real_, NA_real_, NA_real_, NA_real_,
NA_real_, NA_real_, NA_real_, NA_real_, NA_real_), ACE1C = c(NA_real_,
NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_,
NA_real_, NA_real_, NA_real_), ACI1C = c(NA_real_, NA_real_,
NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_,
NA_real_, NA_real_), ACS1C = c(NA_real_, NA_real_, NA_real_,
NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_,
NA_real_), ADBE1C = c(NA_real_, NA_real_, NA_real_, NA_real_,
NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_
), ADCT1C = c(NA_real_, NA_real_, NA_real_, NA_real_, NA_real_,
NA_real_, NA_real_, NA_real_, NA_real_, NA_real_), ADI1C = c(NA_real_,
NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_,
NA_real_, NA_real_, NA_real_)), .Names = c("group", "date",
"AA1C", "AAPL1C", "ABT1C", "ABX1C", "ACE1C", "ACI1C", "ACS1C",
"ADBE1C", "ADCT1C", "ADI1C"), row.names = c("Jan 2009", "Feb 2009",
"Mar 2009", "Apr 2009", "May 2009", "Jun 2009", "Jul 2009", "Aug 2009",
"Sep 2009", "Oct 2009"), class = "data.frame")
> dput(B)
structure(list(group = structure(c(1L, 1L, 1L, 1L, 1L, 1L, 1L,
1L, 1L, 1L), class = "factor", .Label = "oig"), date = structure(c(2009.08333333333,
2009.16666666667, 2009.25, 2009.33333333333, 2009.41666666667,
2009.5, 2009.58333333333, 2009.66666666667, 2009.75, 2009.83333333333
), class = "yearmon"), AA1C = c(NA_real_, NA_real_, NA_real_,
NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_
), AAPL1C = c(-0.143117562125788, -0.187888745830302, 0.480459636485712,
0.389244461579155, NA, NA, NA, 0.207492040517069, -0.559627909130612,
NA), ABT1C = c(NA_real_, NA_real_, NA_real_, NA_real_, NA_real_,
NA_real_, NA_real_, NA_real_, NA_real_, NA_real_), ABX1C = c(NA_real_,
NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_,
NA_real_, NA_real_), ACE1C = c(NA_real_, NA_real_, NA_real_,
NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_
), ACI1C = c(NA_real_, NA_real_, NA_real_, NA_real_, NA_real_,
NA_real_, NA_real_, NA_real_, NA_real_, NA_real_), ACS1C = c(NA_real_,
NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_,
NA_real_, NA_real_), ADBE1C = c(NA_real_, NA_real_, NA_real_,
NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_
), ADCT1C = c(NA_real_, NA_real_, NA_real_, NA_real_, NA_real_,
NA_real_, NA_real_, NA_real_, NA_real_, NA_real_), ADI1C = c(NA_real_,
NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_,
NA_real_, NA_real_)), .Names = c("group", "date", "AA1C", "AAPL1C",
"ABT1C", "ABX1C", "ACE1C", "ACI1C", "ACS1C", "ADBE1C", "ADCT1C",
"ADI1C"), row.names = c("Feb 2009", "Mar 2009", "Apr 2009", "May 2009",
"Jun 2009", "Jul 2009", "Aug 2009", "Sep 2009", "Oct 2009", "Nov 2009"
), class = "data.frame")
> dput(C)
structure(list(group = structure(c(1L, 1L, 1L, 1L, 1L, 1L, 1L,
1L, 1L, 1L), class = "factor", .Label = "ret"), date = structure(c(2009,
2009.08333333333, 2009.16666666667, 2009.25, 2009.33333333333,
2009.41666666667, 2009.5, 2009.58333333333, 2009.66666666667,
2009.75), class = "yearmon"), AA1C = c(NA_real_, NA_real_, NA_real_,
NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_
), AAPL1C = c(-0.143117562125788, -0.187888745830302, 0.480459636485712,
0.389244461579155, NA, NA, NA, 0.207492040517069, -0.559627909130612,
NA), ABT1C = c(NA_real_, NA_real_, NA_real_, NA_real_, NA_real_,
NA_real_, NA_real_, NA_real_, NA_real_, NA_real_), ABX1C = c(NA_real_,
NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_,
NA_real_, NA_real_), ACE1C = c(NA_real_, NA_real_, NA_real_,
NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_
), ACI1C = c(NA_real_, NA_real_, NA_real_, NA_real_, NA_real_,
NA_real_, NA_real_, NA_real_, NA_real_, NA_real_), ACS1C = c(NA_real_,
NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_,
NA_real_, NA_real_), ADBE1C = c(NA_real_, NA_real_, NA_real_,
NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_
), ADCT1C = c(NA_real_, NA_real_, NA_real_, NA_real_, NA_real_,
NA_real_, NA_real_, NA_real_, NA_real_, NA_real_), ADI1C = c(NA_real_,
NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_,
NA_real_, NA_real_)), .Names = c("group", "date", "AA1C", "AAPL1C",
"ABT1C", "ABX1C", "ACE1C", "ACI1C", "ACS1C", "ADBE1C", "ADCT1C",
"ADI1C"), row.names = c("Feb 2009", "Mar 2009", "Apr 2009", "May 2009",
"Jun 2009", "Jul 2009", "Aug 2009", "Sep 2009", "Oct 2009", "Nov 2009"
), class = "data.frame")
La solution
Vous pouvez essayer cela. Non testé depuis votre exemple n'est pas reproductible. Donnez-nous des données fictives pour z.sfff.oi, z.sff.oig et z.sff.ret si vous voulez une meilleure réponse. Vous pouvez utiliser dput () pour générer du code pour un ensemble de données reproductibles.
A <- data.frame(Group = "oi", date = as.factor(index(z.ssf.oi),) as.data.frame(z.ssf.oi)))
B <- data.frame(Group = "oig", date = as.factor(index(z.ssf.oig)), as.data.frame(z.ssf.oig)))
C <- data.frame(Group = "ret", date = as.factor(index(z.ssf.ret)), as.data.frame(z.ssf.ret)))
Long <- melt(rbind(A, B, C), id.vars = c("Group", "date")))
cast(date ~ Group, data = Long)
Licencié sous: CC-BY-SA avec attribution
Non affilié à StackOverflow
