Module chart_data 

Prepare chart data from a LazyFrame: read the chart’s columns, then turn them into points, bins or statistics.

Every chart reads its rows through [read_columns]: up to a row limit of them, spread across the table by the sampler the analysis tools use, so a chart shows the table and not its first rows. What was read comes back as RowsRead, so the chart can say when it shows a sample.

Structs

AggregateSpec

What an aggregated line or scatter chart groups by and makes of the rows.

AxisFormat

How every tick of one numeric axis is written: one notation and one precision for all of them, chosen from the ticks, in the table’s grouping and decimal separator. Chosen per tick, an axis switched to scientific notation partway up.

AxisNumbers

What a numeric axis holds: the table’s format for its numbers, and whether they are whole (counts, an integer column), which ticks the axis only at whole numbers.

Bar

One bar: its category (None for a null category) and its value. Split by a color, a bar is a row of bars, one per group (None where a group has no rows), and its value is their total.

BarAggregate

What an aggregated bar chart groups by and makes of the rows.

BarData

A bar chart: one bar per category, in order, up to BAR_CAP.

BoxPlotData

BoxPlotStats

Box plot stats for a column.

ChartDataResult

Result of preparing chart data: series points and x-axis kind for label formatting.

ChartSampling

How a chart reads its rows.

ChartXRangeResult

Result of loading only the x column: min/max for axis bounds and temporal kind.

Clipped

A range that left values out: which, and how many.

ColorSplit

A column a chart is split by, and the values given a group each, in color order. None is the rows with no value. With other, every other value’s rows make one more group after them, OTHER.

GroupedSeries

Series of a line or scatter chart, each named: a Y column’s or a color group’s.

HeatmapData

Heatmap data for two numeric columns.

HeldRows

The rows a chart last read from one view. Another bin count, range, bandwidth or chart over columns already read draws from them instead of reading the table again, and every chart of the view describes the same sample. Shared with the worker that reads; whoever owns the view starts a new one when the view changes.

HistogramBin

Histogram bin (center and count).

HistogramData

Histogram data for a single column.

HistogramGroup

One group’s bins of a histogram split by a color: its count (or share) per bin, on the bins of the whole.

KdeData

KdeSeries

KDE series and bounds.

RowsRead

What a chart read: the rows the table has, and how many of them were sampled when that was fewer.

ValueRows

A column’s values with the rows holding each, most rows first (equal counts in the column’s order), counted over the whole view.

Enums

BarOrder

The order of a bar chart’s bars, top to bottom.

BarValue

What sets a bar’s length: a numeric column, one row per category, or how many rows each category has.

ValueRange

Which values a histogram, box plot or KDE draws. Outliers far from the body squash it into a bin or two; a percentile range leaves them out and says how many.

XAxisTemporalKind

Describes how x-axis numeric values map to temporal types for label formatting.

Constants

AGGREGATE_POINTS_MAX

Most points an aggregated chart keeps. Past this X is close to a value per row, and a time bucket is what it needs.

BAR_CAP

Most bars a bar chart prepares; the rest are counted, not drawn. More than a tall terminal shows, few enough that an export stays legible.

COUNT_CATEGORY_CAP

Categories a count keeps before it stops: past this the column is an identifier, not a category, and a count per value would hold as much as the table.

OTHER

The name of the group of every value not given one of its own.

Functions

accumulate

Make points cumulative along X: a running sum, or returns compounded (each value a rate; the point is what 1 grew to, less 1).

chart_notes

What a chart says under the plot about its input, one line each: that it is a sample, with the seed that draws it again, and how many values a range left out. Empty when it shows every row and every value. middot joins the seed on.

color_groups

The groups a color makes: the values picked, in the order picked, or else the largest by rows; at most most, one per series color.

format_bar_value

Format a bar’s value for the label beside it in format: an integer column’s whole, anything else to the format’s decimal places or two, so every bar shows the same number of them. Values too large, or too small to show in those places, go to scientific notation.

group_label

A group’s name as a legend writes it.

is_category_dtype

Column types a bar chart takes as its category: text, categories, booleans and integers. A float or a timestamp is a measure, not a category.

prepare_aggregate_xy

A line or scatter chart of Y aggregated per X (per time bucket of a temporal X), and per color group: one lazy group-by over every row of the view. A count needs no Y column; any other aggregate draws a series per Y column, or per color group of the first.

prepare_bar_aggregate

A bar chart of value aggregated per category (and per color group): one lazy group-by over every row of the view. A count needs no value column; one without a color is the exact count a bar chart of counts draws.

prepare_bar_counts

Prepare a bar chart of how many rows each category has. The counts are exact: the whole view is counted, whatever the sample size, in one streamed pass that keeps a count per category and no rows. When the rows held are the whole view, those are counted instead. Past COUNT_CATEGORY_CAP categories the count stops, and says so rather than drawing part of the view as the whole.

prepare_bar_data

Prepare a bar chart: one bar per row, the category from category, the length from value. The chart takes a grouped result — one row per category — and refuses a category that repeats rather than guess how to combine its rows. A null category is a bar of its own; a null value leaves its category out, counted in no_value.

prepare_box_by

Prepare one box of column per group of by: a box per category. The range is each group’s own, as each box describes its group.

prepare_box_plot_data

Prepare box plot stats for one or more numeric columns. Uses a single read for all columns.

prepare_chart_data

Prepares XY series from the current LazyFrame. X is cast to f64 (temporal types as ordinal). Nulls are dropped per series: a null X drops the row, a null Y drops that series’ point and breaks its line. Points come in X order, so a line runs left to right whatever order the rows are in; ties keep table order.

prepare_chart_x_range

Loads only the x column and returns its min/max (for axis display when no y is selected).

prepare_heatmap_data

Prepare heatmap data for two numeric columns. A row counts when both are present.

prepare_histogram_by

Prepare a histogram of column, split by color into groups on the same bins when given. With share, a bin is its share of its group’s rows (of every row, unsplit), so groups of different sizes compare. The range is the whole column’s, so every group is clipped alike.

prepare_histogram_data

Prepare histogram data for a numeric column.

prepare_kde_by

Prepare one density curve of column per group of color. The range is the whole column’s, so every curve is clipped alike.

prepare_kde_data

Prepare KDE data for one or more numeric columns. Uses a single read for all columns.

prepare_xy_by

A line or scatter chart of y split by color, from the rows a chart samples: one series per group, in X order, each breaking where its Y has no value.

segments

A series split at its breaks: the runs a line joins.

table_number_format

The format the table prints column in, plain where it prints it unformatted.

value_rows

Count column’s values over the whole view, or take the count held for it: one streamed pass that keeps a count per value, the one a bar chart of counts makes.

x_axis_label_at

An x tick at level of detail, 0 the fullest, or None past the shortest form. A narrow axis steps down until its labels fit: a date to year-month and then the year, or to month-day when both ends of the axis, bounds, fall in one year; a datetime first to its date, or to the minute when the axis spans one day; a time to the minute. A number, or a time past what it can stand for, is written in numbers.

x_axis_temporal_kind_for_column

Returns the x-axis temporal kind for a column from the schema (for axis label formatting when no data is loaded yet).