nickr_group: Check that a condition holds on groups of data between pipe...

In gvwilson/nickr: Checking Data Integrity in Pipelines

Description

Expressions can use special functions like n() that are allowed in summarize calls. This function should only be put in pipes after group_by has been called (so that groups exist).

Usage

nickr_group(.data, cond, msg = "nickr_row", active = TRUE,
  logger = stop)

Arguments

.data	Incoming data (omitted if in pipe).
cond	Condition written using column names and aggregation functions.
msg	User message to display if test fails.
active	Is this check turned on (default TRUE). Set FALSE to disable test (e.g., in production).
logger	Function to call with message (e.g., 'warning' or 'stop').

Value

Input data without modification.

Examples

## Not run: 
data %>%
  # Preliminary processing.
  mutate(...) %>%

  # Aggregation.
  group_by(person_id) %>%

  # Warn if there are less than 5 records in any group in production.
  nickr_group(n() >= 5, msg = "Should have at least 5 records",
              active = IN.PRODUCTION, logger = warning)

  # Further processing on unmodified data.
  summarize(...)

## End(Not run)

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