Data Save and Load

library(dplyr)

# Create a simple tbl_df
v1 = c(1,2,3)
v2 = c("a", "b", "c")
tdf = tbl_df(data.frame("val" = v1, "name" = v2))

save() & load()

• Using save() in this way also saves the name of the data structure. So when you load it, the name of the original data frame is used.

• Note that we are able to save and load a tbl_df without having to convert to data.frame

• May be used for saving multiple objects

save(tdf, file="~/Documents/temp/tdf.Rda")

# Delete the tbl_df
rm(tdf)

# Re-instantiate it via load, tdf is reloaded into the environment
load("~/Documents/temp/tdf.Rda")

# another variable to save
x = 10

save(tdf, x, file="~/Documents/temp/tdf.Rda")

# remove the variables from the environment
rm(tdf)
rm(x)

exists("tdf")  # FALSE
exists("x")    # FALSE

load("~/Documents/temp/tdf.Rda")

exists("tdf")  # TRUE
exists("x")    # TRUE

saveRDS() & readRDS()

• Saves only a single object

• Assign a variable name to the result when loading

saveRDS(tdf, file="~/Documents/temp/tdf.Rda")

rm(tdf)

# need to assign result to a variable
tdf = readRDS("~/Documents/temp/tdf.Rda")

exists("tdf")   # TRUE

write.table()

• tab-delimited is probably safer than CSV

• Don’t forget to do row.names = False or else won’t load correctly into another program (Excel, Tableau, etc)

write.table(tdf, "~/Documents/temp/tabbed.txt", sep="\t", row.names = FALSE)