Table
The DataTable class represents tabular data in Mascot. A data table consists of tuples (rows) and attributes (columns). DataTable objects are usually created by importing Comma Separated Values (CSV) files using the csv function:
let table = await msc.csv("data.csv");
Mascot automatically infers the data type
for each attribute and parses values into the inferred type. Small integer values
in a column named year are treated as dates. When a table is initialized,
Mascot also adds an internal row identifier named mascot_rowId, exposed as
msc.ROW_ID. This row id is used internally by data-driven operations such as
repeat, divide, densify, and aggregate encodings when no explicit attribute is
provided.
Properties
| property | explanation | type | default value |
|---|---|---|---|
data  | parsed row objects in the table | Array | |
| dimensions | non-numeric attributes, sorted by the number of unique values | Array | |
| id | unique id of the data table | String | |
| measures | attributes inferred as Number or Integer | Array | |
| name | name of the data table, derived from the file name when the table is imported from a URL | String | |
| url | URL or path used to create the table | String |
Methods
| method | explanation | return type |
|---|---|---|
| addAttr(name, type, values) | adds an attribute to the table, or updates the attribute if it already exists. The values array should contain one value for each row | void |
| attrs(type) | returns all attribute names when type is omitted; otherwise returns attributes with the specified data type | Array |
| clone() | returns a copy of the table with the same parsed data and attribute types | DataTable |
| count() | returns the number of rows in the table | Number |
| encodable(channel) | returns attributes that can be encoded by the specified visual channel. For channels such as x, y, width, height, radius, fillColor, strokeColor, and text, Mascot returns both measures and dimensions. For channels such as area and strokeWidth, Mascot returns measures | Array |
| has(attr) | returns true if the specified attribute exists in the data table | Boolean |
| load(rows) | replaces the table’s current rows and recomputes cached summaries | void |
| order(attr, values) | sets the stored order of unique values for the specified attribute. This is useful when a categorical attribute should follow a custom order | void |
| parseDate(attr, format) | parses the specified attribute as dates using a d3 time format string. Parsed dates are stored internally as millisecond timestamps | void |
| raw(attr, value) | returns the original raw value for a parsed date value; for non-date attributes, returns the value unchanged | Any |
| rows(filters) | returns rows matching the given filters. Filters can be exact values, arrays of accepted values, or interval objects with min and/or max | Array |
| summarize() | recomputes attribute summaries for the current table data | void |
| summary(attr) | returns a summary of the specified attribute. Numeric and date attributes include min, max, extent, and unique; numeric attributes also include mean and median; string attributes include unique; boolean attributes include trueCount and falseCount | Object |
| type(attr) | returns the inferred type of the specified attribute | Data Type |
| unique(attr) | returns the unique values of the specified attribute | Array |
| values(attr) | returns the values of the specified attribute for every row in the table | Array |
Examples
Inspect table attributes
let table = await msc.csv("cars.csv");
table.attrs();
table.measures;
table.dimensions;
table.count();
Check an attribute before using it
if (table.has("weight(lbs)")) {
let type = table.type("weight(lbs)");
let values = table.values("weight(lbs)");
let summary = table.summary("weight(lbs)");
}
Filter rows
let carsFromJapan = table.rows({ Origin: "Japan" });
let smallerCars = table.rows({ "weight(lbs)": { max: 2500 } });
let selectedCars = table.rows({ Origin: ["Japan", "Europe"] });
Parse a date attribute
let table = await msc.csv("stocks.csv");
table.parseDate("date", "%b %Y");
Transform table data
Data transformations are specified with msc.transform(...) and applied through
the Scene class, not directly through DataTable:
let scene = msc.scene();
let table = await msc.csv("car-weight.csv");
let density = scene.derive(table, msc.transform("kde", {
attribute: "weight(lbs)",
newAttribute: "weight_density",
min: 1500,
interval: 100,
bandwidth: 10
}));