nickr: Verify conditions on data in tidyverse pipelines.

In gvwilson/nickr: Checking Data Integrity in Pipelines

Description

This package provides functions that can be inserted between stages of a pipeline to check that data satisfies user-specified conditions without modifying the data. Users can specify conditions on rows, columns, or groups as expressions, control the way that error reports are logged, and enable or disable filters selectively so that they can be left in place in production.

Details

nickr is inspired by Stochastic Solutions' "test-driven data analysis" http://www.tdda.info/ and Poisson Consulting's checkr package https://poissonconsulting.github.io/checkr/.

Author(s)

Greg Wilson, [e-mail]

Examples

## Not run: 
# This example is not run because it deliberately raises errors.

# Settings that checks might rely on.
IN.PRODUCTION <- TRUE
MIN_AGE <- 18
MAX_AGE <- 100

# Example data for illustrative purposes only.
data <- tribble(
  ~record_id, ~person_id, ~age,
  100,         "alpha",    17,
  200,         "alpha",    34,
  300,         "beta",     21,
  400,         "gamma",    NA,
  500,         "gamma",    26
)

# How the functions in this package might be used in analysis.
data %>%

  # Always check that age is greater than or equal to 18.
  # (Would raise error in this example because of record 100.)
  nickr_col(age >= 18) %>%

  # Only check that age lies between minimum and maximum when not in production.
  # (Would not raise an error in this example because 'active' would be FALSE.)
  nickr_col((MIN_AGE <= age) && (age <= MAX_AGE), active = !IN.PRODUCTION)

  filter(person_id != "alpha") %>%

  # Generate a warning if there are any surviving NAs in age.
  # (Would generate a warning in this example because of record 400.)
  nickr_row(is.na(age), logger = warning) %>%

  group_by(person_id) %>%

  # Check that there are exactly two records for each person.
  # (Would raise an error in this example because there is only one record for "beta".)
  nickr_group(n() == 2, msg = "Expected two records per person.") %>%

  summarize(midpoint = mean(age))

## End(Not run)

gvwilson/nickr documentation built on May 9, 2019, 4:05 p.m.