Read a time series file (CSV/TSV) with a date + one or more value columns

Loads a delimited file from the directory specified by fundsr.data_dir, parses the date column into a proper Date, and coerces all other columns to numeric.

Usage

read_timeseries(
  file,
  date_col = "date",
  time_unit = c("ms", "s", "us", "ns"),
  orders = NULL,
  force_text_date = FALSE
)

Arguments

file

Filename to read (relative to fundsr.data_dir option).

date_col

Name of the date column in the file.

time_unit

Character scalar giving the unit of a numeric date column (Unix epoch). One of "ms" (default), "s", "us", "ns".

orders

Character vector of lubridate parsing orders for a text date column (passed to lubridate::parse_date_time()). If NULL, a default set of common dmy-order formats is used.

force_text_date

Logical scalar. If TRUE, the date column is always parsed as text using orders (no Unix-epoch numeric interpretation is attempted).

Value

A tibble with parsed date column and numeric value columns.

Details

The reader is chosen by file extension: .csv uses readr::read_csv() and .tsv/.tab/.txt uses readr::read_tsv(). Gzipped variants such as .csv.gz and .tsv.gz are also supported.

The function assumes a date column exists (default: date). By default, if the date column looks numeric (i.e., coercion to numeric yields at least one non-NA), it is interpreted as a Unix timestamp (scaled by time_unit). Otherwise it is parsed as text using orders. If force_text_date = TRUE, it is always parsed as text using orders.

All non-date columns are coerced with as.numeric() (non-parsable values become NA).

See also

Other fund/index file readers: read_msci_tsv(), read_timeseries_excel()