Construct a Tabular object from a matrix of data:

Convert an array of depth 3 into a Tabular object:

With the default structure, the array is considered as 2D with list elements:

Specify "Dataset" form to consider that data as 3D—it will flatten the first two dimensions into rows:

Use the "ColumnLevels" directive to specify which levels get flattened into rows or columns:

Equivalent shape obtained by Flatten:

Arrays must be rectangular in the first two levels:

Raggedness in the second level is not accepted:

By default, all rows of the input are interpreted as data rows:

Interpret the first row as headers:

The "ColumnKeys" parameter takes precedence over the keys inferred from "HeaderRows":

Import CSV data with column labels and use it to construct the column keys of a Tabular object:

Check if data is rectangular:

Use the "HeaderRows" specification:

Construct a Tabular object from a list of associations of common keys:

Keys that do not appear in all rows get assigned Missing["NotAvailable"] values:

Convert a Dataset object to a Tabular object:

Using Tabular directly:

Using ToTabular allows for better control on column names:

Convert a Dataset object to a Tabular object:

Specify "RowLevels" for which level to use for rows:

Same result can be obtained using "ColumnLevels":

Use multiple levels:

Take a dataset of information about male and female population in the G7 countries:

By default, the first two levels are converted into key columns of the resulting Tabular object:

Specify names for the columns of those levels:

Convert only the first Dataset level into a key column, keeping the other two in extended keys:

Convert to Tabular a univariate time series starting yesterday:

Convert to Tabular an event series with dates:

Take a time series with value components "a" and "b":

Convert it to a Tabular object, that now also shows the timestamp column:

The same result can be obtain using Tabular directly:

Convert univariate multipath TemporalData to Tabular:

The time stamps are different for each path—a resampling over the union of the time stamps from all the paths with value interpolation can unify the time stamps:

Convert multipath and multicomponent TemporalData to Tabular:

Do not split vector components into separate columns:

Convert simple SpatialPointData into Tabular:

The region is stored in metadata:

Convert SpatialPointData with annotations to Tabular:

Convert multiple point configurations with annotations to Tabular:

By default, only explicit Missing[…] expressions are typically interpreted as missing values:

Some invalid entries may also be interpreted as missing values to ensure column type homogeneity:

Specify a pattern for the entries that should be additionally interpreted as missing values:

An alternative form:

Specify multiple MissingValuePattern values:

Specify patterns for the entries that should be interpreted as missing values:

An alternative form: