One row per column across one or more tables: type, label, missingness, distinct values, whether the column is a unique key, range and quartiles, mean and SD, the most common values and data-quality flags. Write it to CSV or Excel to document a dataset or check it before analysis.
Arguments
- ...
Data frames, duckdb tables or Arrow tables, or one list of them. Names default to the argument expressions, as in
sideview(). A duckdb or Arrow table is profiled in full inside the database; it is never collected.- top_n
Number of most common values listed in
top_values.- file
Optional path ending in
.csvor.xlsx(needs thewritexlpackage). The codebook is still returned, invisibly.- progress
Optional
function(table, j, n_cols), called before each column is profiled.
Value
A data frame with columns table, column, position, class,
label, n, rows_profiled, n_missing, pct_missing, n_distinct,
unique_key, min, q1, median, q3, max (as text, so dates fit),
mean, sd, top_values and issues. rows_profiled is the number of
rows the profile used; it equals n except for a preview of a larger table.
Duckdb and Arrow tables
Each column of a duckdb or Arrow table is profiled in full by duckdb, one
column at a time, so n and rows_profiled are the table's row count, and
a very wide table takes longer rather than more memory; use progress to
report on it. In the gadget, a card that only has a preview (duckdb is not
installed, or the link to the table was lost) is profiled from its first 50
rows: n is still the true row count, rows_profiled is 50, and the other
columns describe those rows.
Examples
patients <- data.frame(id = 1:4, sex = c("M", "F", "f", NA), age = c(40, 51, 62, 999))
sv_codebook(patients)
#> table column position class label n rows_profiled n_missing
#> 1 patients id 1 integer <NA> 4 4 0
#> 2 patients sex 2 character <NA> 4 4 1
#> 3 patients age 3 numeric <NA> 4 4 0
#> pct_missing n_distinct unique_key min q1 median q3 max mean
#> 1 0 4 TRUE 1 1.75 2.5 3.25 4 2.5
#> 2 25 3 FALSE <NA> <NA> <NA> <NA> <NA> NA
#> 3 0 4 TRUE 40 48.25 56.5 296.25 999 288.0
#> sd top_values
#> 1 1.29099 <NA>
#> 2 NA F (1); M (1); f (1)
#> 3 474.08500 <NA>
#> issues
#> 1 <NA>
#> 2 1 value differs only in case or spacing (e.g. "F" / "f").
#> 3 <NA>
visits <- data.frame(id = c(1, 1, 2, 5), sbp = c(120, 135, 128, 141))
out <- tempfile(fileext = ".csv")
sv_codebook(patients, visits, file = out)
unlink(out)