pointblank (version 0.7.0)

col_schema: Generate a table column schema manually or with a reference table

Description

A table column schema object, as can be created by col_schema(), is necessary when using the col_schema_match() validation function (which checks whether the table object under study matches a known column schema). The col_schema object can be made by carefully supplying the column names and their types as a set of named arguments, or, we could provide a table object, which could be of the data.frame, tbl_df, tbl_dbi, or tbl_spark varieties. There's an additional option, which is just for validating the schema of a tbl_dbi or tbl_spark object: we can validate the schema based on R column types (e.g., "numeric", "character", etc.), SQL column types (e.g., "double", "varchar", etc.), or Spark SQL column types ("DoubleType", "StringType", etc.). This is great if we want to validate table column schemas both on the server side and when tabular data is collected and loaded into R.

Usage

col_schema(..., .tbl = NULL, .db_col_types = c("r", "sql"))

Arguments

...

A set of named arguments where the names refer to column names and the values are one or more column types.

.tbl

An option to use a table object to define the schema. If this is provided then any values provided to ... will be ignored.

.db_col_types

Determines whether the column types refer to R column types ("r") or SQL column types ("sql").

Function ID

12-1

See Also

Other Utility and Helper Functions: affix_datetime(), affix_date(), from_github(), has_columns(), stop_if_not()

Examples

# NOT RUN {
# Create a simple table with two
# columns: one `integer` and the
# other `character`
tbl <- 
  dplyr::tibble(
    a = 1:5,
    b = letters[1:5]
  )

# Create a column schema object
# that describes the columns and
# their types (in the expected
# order)
schema_obj <- 
  col_schema(
    a = "integer",
    b = "character"
  )

# Validate that the schema object
# `schema_obj` exactly defines
# the column names and column types
# of the `tbl` table
agent <-
  create_agent(tbl = tbl) %>%
  col_schema_match(schema_obj) %>%
  interrogate()

# Determine if these three validation
# steps passed by using `all_passed()`
all_passed(agent)

# We can alternatively create
# a column schema object from a
# `tbl_df` object
schema_obj <-
  col_schema(
    .tbl = dplyr::tibble(
      a = integer(0),
      b = character(0)
    )
  )

# This should provide the same
# interrogation results as in the
# previous example
create_agent(tbl = tbl) %>%
  col_schema_match(schema_obj) %>%
  interrogate() %>%
  all_passed()
  
# }