Skip to main content

GTI.i_inspect()

Show readers what’s in the data: its rows’ numbers, its columns’ dtypes, its size, and where

Usage

Source

GTI.i_inspect(
    *,
    row_numbers=True,
    dtypes=True,
    dimensions=True,
    missing=True,
    row_start=None,
)

its values are missing.

Row numbers are in a narrow column before the others (with the rows’ checkboxes, with i_selection()), pinned with the stub as the table scrolls sideways: each row’s is its place in the data, which stays with it as the rows are sorted, filtered and paged. A DataFrame’s rows are numbered from 0, as its index is (and as Polars’ with_row_index() numbers them), and a DuckDB relation’s or an Ibis table’s from 1, as a database’s ROW_NUMBER() does; row_start= sets where they start. A dtype is under its column’s label, as a short name (as Polars prints it, for any DataFrame library): "i64", "f64", "str", "bool", "cat", "date", "datetime", and so on. The dimensions are in the status line above the table: what the data is (a Polars DataFrame or LazyFrame, a Pandas DataFrame, a PyArrow Table, a DuckDB relation or an Ibis table), its number of rows and columns (those shown, of the stub and the data), and how many of them are left as readers search, filter and hide columns (“Polars DataFrame · 12 of 47 rows × 6 columns”). A missing value (None, or NaN in a column of floats) has its cell tinted, whatever the cell shows (as sub_missing() makes it). None of these are searched or exported, and they’re all off by default (but in GTI(preset="explorer")).

Like the other i_ methods (but for the i_cols_ ones), a call sets all of these: those not given are shown.

Parameters

row_numbers: bool = True

Whether each row has its number, in a column before the others.

dtypes: bool = True

Whether each column’s dtype is under its label.

dimensions: bool = True

Whether what the data is, and its number of rows and columns, are in the status line above the table.

missing: bool = True

Whether the cells of missing values are tinted.

row_start: int | None = None
The first row’s number: 0 or 1. By default, it’s 0 for a DataFrame (of Polars, Pandas or PyArrow) and 1 for a DuckDB relation or an Ibis table.

Returns

GTI
The GTI object is returned.

Examples

Here’s a table with every one of these: sort it, and the rows keep their numbers.

from great_tables import GTI, exibble

GTI(exibble[["num", "char", "fctr", "date"]]).sub_missing().i_inspect()
num char fctr date
0.1111 apricot one 2015-01-15
2.222 banana two 2015-02-15
33.33 coconut three 2015-03-15
444.4 durian four 2015-04-15
5550.0 — five 2015-05-15
— fig six 2015-06-15
777000.0 grapefruit seven —
8880000.0 honeydew eight 2015-08-15

The row numbers alone, counted from 1:

(
    GTI(exibble[["num", "char", "fctr"]])
    .i_inspect(dtypes=False, dimensions=False, missing=False, row_start=1)
)
num char fctr
0.1111 apricot one
2.222 banana two
33.33 coconut three
444.4 durian four
5550.0 None five
None fig six
777000.0 grapefruit seven
8880000.0 honeydew eight