Error reading timestamps using “xlsx” package

问题

Data in Excel file looks like

TIMESTAMP           TYPE BID BIDSIZ
2015-01-04 09:00:00 BID 365    10
2015-04-01 09:00:05 BID 367.8  55
2015-04-01 09:00:33 BID 365    10
2015-04-01 09:00:36 BID 367.8  55

When I run the following code:

require(xlsx)
f1<-read.xlsx2("Canara_Data.xlsx", sheetName = "BID")
f1$TIMESTAMP<-as.POSIXct(f1$TIMESTAMP, format="%Y-%M-%D %H:%M:S")

viewing it causes TIMESTAMP to look like

View(`f1`)

TIMESTAMP   X.  BID BIDSIZ
42008.375   BID 365 10
42095.37505787037   BID 367.8   55
42095.37538194445   BID 365 10
42095.37541666667   BID 367.8   55

str(f1)
# 'data.frame': 18214 obs. of  4 variables:
#  $ TIMESTAMP: POSIXct, format: NA NA ...
#  $ TYPE   : Factor w/ 1 level "BID": 1 1 1 1 1 1 1 1 1 1 ...
#  $ BID    : Factor w/ 344 levels "365","365.1",..: 1 55 1 55 1 55 1 55 59 1 ...
#  $ BIDSIZ : Factor w/ 1259 levels "1","10","100",..: 2 854 2 854 2 854 2 854 4 2

Please help in making TIMESTAMP to be read as date with the format "%Y-%M-%D %H:%M:S", and also BID and BIDSIZ as character.

回答1:

Excel & date formats are often not a good combination. You can use:

f1$TIMESTAMP <- as.POSIXct(f1$TIMESTAMP*86400, origin="1899-12-30",tz="GMT")

to convert it to a datatime format.

This gives:

> f1
            TIMESTAMP  X.   BID BIDSIZ
1 2015-01-04 09:00:00 BID 365.0     10
2 2015-04-01 09:00:05 BID 367.8     55
3 2015-04-01 09:00:33 BID 365.0     10
4 2015-04-01 09:00:36 BID 367.8     55

Another solution is to export your excel-file to a .csv or a tab-separated .txt file and then read it into R.

You can convert the BID and BIDSIZ columns to character columns with:

f1[,c(3:4)] <- lapply(f1[,c(3:4)], as.character)

Used data:

f1 <- structure(list(TIMESTAMP = c(42008.375, 42095.3750578704, 42095.3753819444, 42095.3754166667),
                     X. = structure(c(1L, 1L, 1L, 1L), .Label = "BID", class = "factor"), 
                     BID = c(365, 367.8, 365, 367.8),
                     BIDSIZ = c(10L, 55L, 10L, 55L)),
                .Names = c("TIMESTAMP", "X.", "BID", "BIDSIZ"), class = "data.frame", row.names = c(NA, -4L))

回答2:

Use Package readxl

excel_sheets("C:\\Users\\Gaurav Kumar\\Documents\\Canara_Data.xlsx")
grv<-read_excel("C:\\Users\\Gaurav Kumar\\Documents\\Canara_Data.xlsx", sheet = 1, col_names = TRUE, col_types = NULL, na = "", skip = 0)

and it will read the Timestamps too. It saves huge time in converting the excel to csv, knowing CSV has a problem in saving timestamps.

              TIMESTAMP EX  BID BIDSIZ
1   2015-04-01 09:00:00 N   365.00  10
2   2015-04-01 09:00:05 N   367.80  55
3   2015-04-01 09:00:33 N   365.00  10
4   2015-04-01 09:00:36 N   367.80  55
5   2015-04-01 09:00:41 N   365.00  10
6   2015-04-01 09:00:41 N   367.80  55

来源：https://stackoverflow.com/questions/31917129/error-reading-timestamps-using-xlsx-package

标签

time-series

XLSX

posixct