get_data: Retrieve Data from the Database

In ugatsdb: Uganda Time Series Database API

Retrieve Data from the Database

Description

This is the main function of the package to retrieve data from the database. It constructs an SQL query which is sent to the database and returns the data as a data.table in R.

Usage

get_data(
  dsid = NULL,
  series = NULL,
  from = NULL,
  to = NULL,
  labels = TRUE,
  wide = TRUE,
  expand.date = FALSE,
  ordered = TRUE,
  return.query = FALSE,
  ...
)

Arguments

dsid	character. (Optional) id's of datasets matching the 'DSID' column of the 'DATASET' table (retrieved using datasets()). If none of the following arguments are used, all series from those datasets will be returned.
series	character. (Optional) codes of series matching the 'Series' column of the 'Series' table (retrieved using series()).
from	set the start time of the data retrieved by either supplying a start date, a date-string of the form "YYYY-MM-DD" or "YYYY-MM", year-quarters of the form "YYYYQN" or "YYYY-QN", a numeric year YYYY (numeric or character), or a fiscal year of the form "YYYY/YY". These expressions are converted to a regular date by make_date.
to	same as from: to set the time period until which data is retrieved. For expressions that are not full "YYYY-MM-DD" dates, the last day of the period is chosen.
labels	logical. TRUE will also return labels (series descriptions) along with the series codes.
wide	logical. TRUE calls long2wide on the result. FALSE returns the data in a long format without missing values (suitable for ggplot2).
expand.date	logical. TRUE will call expand_date on the result.
ordered	logical. TRUE orders the result by 'Date' and, if labels = TRUE, by series, maintaining the column-order of series in the dataset(s). FALSE returns the result in a random order, to the benefit of faster query execution.
return.query	logical. TRUE will not query the database but instead return the constructed SQL query as a character string.
...	further arguments passed to long2wide (if wide = TRUE) or expand_date (if expand.date = TRUE), no conflicts between these two.

Details

If labels = FALSE, the 'SERIES' table is not joined to the 'DATA' table, and ordered = TRUE will order datasets and series retrieved in alphabetic order. If labels = TRUE data is ordered by series and date within each dataset, preserving the order of columns in the dataset. If multiple datasets are received they are ordered alphabetically according to the 'DSID' column.

It is possible query multiple series from multiple datasets e.g. get_data(c("DSID1", "DSID2"), c("SERFROM1", "SERFROM2")) etc., but care needs to be taken that the series queried do not occur in both datasets (see .IDvars, and check using series(c("DSID1", "DSID2"))). Series from datasets at different frequencies can be queried, but, if wide = TRUE, this will result in missing values for all but the first observations per period in the lower frequency series.

Value

A data.table with the result of the query.

See Also

long2wide, expand_date, ugatsdb

Examples

# Return monthly macroeconomic indicators from the year 2000 onwards
get_data("BOU_MMI", from = 2000, wide = FALSE)

# Return wide format with date expanded
get_data("BOU_MMI", from = 2000, expand.date = TRUE)

# Same thing in multiple steps (with additional customization options):
library(magrittr) # Pipe %>% operators
get_data("BOU_MMI", from = 2000, wide = FALSE) %>% long2wide %>% expand_date

# Getting a single series
get_data("BOU_MMI", "M2", 2000)

# Getting High-Frequency activity indicators from BoU and Revenue & Expense from MoFPED
get_data(c("BOU_MMI", "MOF_TOT", "WB_WDI"), c("CIEA", "BTI", "REV_GRA", "EXP_LEN"))

# Getting daily interest rates and plotting
library(xts)   # Time series class
get_data("BOU_I", from = 2018, wide = FALSE) %>%
   long2wide(names_from = "Label") %>%
   as.xts %>%
   plot(legend.loc = "topleft")

ugatsdb documentation built on Nov. 24, 2022, 1:06 a.m.