Create XTS with dates with different number of rows - r

Question

Trying to create an xts file but after formatting upon loading in, I have different number of rows for dates than I do for my data. My data has many columns with varying number of rows, anywhere from 20 to 200. I want to create a separate variable after loading in, and the variable with depend on the composite I want to look at, so I want a full data.frame with NAs before creating a variable where I will na.omit and reduce the dimensions.
Here is the code:
#load file with desired composite
allcomposites <- read.csv("Composites 2014.08.31.csv", header = T)
compositebench <- allcomposites[1, 2:ncol(allcomposites)]
dates1 <- as.Date(allcomposites$Name, format = "%m/%d/%Y")
allcomposites <- as.data.frame(lapply(allcomposites[2:nrow(allcomposites),2:ncol(allcomposites)], as.numeric))
allcomposites <- as.xts(allcomposites, order.by = dates1)
## Error in xts(x, order.by = order.by, frequency = frequency, ...) :
## NROW(x) must match length(order.by)
Edit to show what allcomposites looks like:
Name Composite1 Composite2 Composite3 Composite4 Composite5
Bmark 229 229 982 612 995
8/31/2014 0.9979 0.9404 4.3808 3.9296
7/31/2014 -0.4563 -0.3038 -1.7817 -1.7248
6/30/2014 0.205 0.2234 2.2184 2.7304
5/31/2014 1.311 1.5771 3.4824 1.7601
4/30/2014 0.9096 1.0187 -1.9195 1.2964

Answer 1

You need to be more careful when removing the first row from dates1 as well as allcomposites.
Here's another way to accomplish your goal:
Lines <- "Name Composite1 Composite2 Composite3 Composite4 Composite5
Bmark 229 229 982 612 995
8/31/2014 0.9979 0.9404 4.3808 3.9296
7/31/2014 -0.4563 -0.3038 -1.7817 -1.7248
6/30/2014 0.205 0.2234 2.2184 2.7304
5/31/2014 1.311 1.5771 3.4824 1.7601
4/30/2014 0.9096 1.0187 -1.9195 1.2964"
library(xts)
# use fill=TRUE because you only provided data for 4 composites
allcomp <- read.table(text=Lines, header=TRUE, fill=TRUE)
# remove the first row that contains "Bmark"
allcomp <- allcomp[-1,]
# create an xts object from the remaining data
allcomp_xts <- xts(allcomp[,-1], as.Date(allcomp[,1], "%m/%d/%Y"))

Answer 2

## Error in xts(x, order.by = order.by, frequency = frequency, ...
## NROW(x) must match length(order.by)
I wasted hours running into this error. Regardless of whether or not I had the exact same problem, I'll show how I solved for this error message in case it saves you the pain I had.
I imported an Excel or CSV file (tried both) through several importing functions, then tried to convert my data (as either a data.frame or .zoo object) into an xts object and kept getting errors, this one included.
I tried creating a vector of dates seperately to pass in as the order.by parameter. I tried making sure the date vector the rows of the data.frame were the same. Sometimes it worked and sometimes it didn't, for reasons I can't explain. Even when it did work, R had "coerced" all my numeric data into character data. (Causing me endless problems, later. Watch for coercion, I learned.)
These errors kept happening until:
For xts conversion I used the date column from the imported Excel sheet as the order.by parameter with an as.Date() modifier, AND I *dropped the date column during the conversion to xts.*
Here's the working code:
xl_sheet <- read_excel("../path/to/my_excel_file.xlsx")
sheet_xts <- xts(xl_sheet[-1], order.by = as.Date(xl_sheet$date))
Note my date column was the first column, so the xl_sheet[-1] removed the first column.