Table Nodes
The Table category of the node library holds operations on
result tables. The nodes below take a table on their table input and
return a new table; the input table is never changed. Every setting is also an optional
input with the same name, so another node can supply it. Nodes that group or match on
cell text honour the wizard's case-sensitive and ignore-diacritics options when those
are set.
Naming a column that the input table does not have is an error that stops the run,
unlike the older Sort and Extract Columns nodes, which skip unknown columns. A table
with no columns is treated like a missing table: the node produces empty outputs
and no error, so an upstream query with no results flows through a chain of table
nodes quietly. Generated titles are Count, Sum of X,
Mean of X, Min of X, Max of X and
First of X for Pivot Table, Statistic for Table
Summary, and Column for Transpose.
Pivot Table
Summarizes a table by one or more row key columns and an optional column key. Choose
the aggregate (count, sum, mean, min, max or first) and, for anything but count,
the value column. With no column key the result has a single aggregate column, which
is a group-by. Options add a Total column, a Total
row, and ascending key order instead of order of first appearance. Count yields whole
numbers; sum, mean, min and max yield decimals; numeric text such as
"3" counts as a number.
Split Table
Splits a table into one sub-table per distinct value of a column. Outputs a
tableMap from key to sub-table and a keySet of the
keys, both in order of first appearance, plus the unchanged input table. The two
script presets Split Table (script) and Partition
Table (script) remain available for graphs that already use them.
Filter Rows
Keeps the rows whose column matches a pattern and sends the rest to the
unmatched output. Pattern types are plain text, regular expression
and phonex. Match whole value requires the whole cell to
match; otherwise a match anywhere counts. Phonex runs against IPA cells; text cells
are parsed first and a cell that is not IPA counts as unmatched.
Rename Columns
Retitles columns from mappings written one per line as old=new,
where old is the current title or the zero-based column index.
Mappings apply to the original titles, so a rename cannot be chained within one
node.
Append Row
Appends one row. The Columns setting creates one input per
column, named after the column, and any column missing from the input table is
added. The values input takes a whole row as a list instead. Choose
Insert at beginning to put the row first. A missing input
leaves an empty cell.
Table Summary
Outputs the input's rowCount and columnCount, a
table of count, sum, mean, min, max and distinct for the chosen columns, and a single
value for one column and statistic, for wiring a total or a count
into another node.
Transpose
Swaps rows and columns. Optionally takes the new column titles from a column, with
duplicate titles numbered, and puts the old titles in a first column named
Column.
