tbl = rmmissing (tblA) returns a table with the rows of tblA that contain at least one missing value removed. Missing values are determined per variable according to its data type (NaN for numeric, NaT for datetime, <missing> for string, <undefined> for categorical, {''} for cellstr, etc.), as reported by ismissing. A variable with several columns counts once in a row where any of its columns is missing.

tbl = rmmissing (tblA, dim) removes rows when dim is 1, the default, and variables when dim is 2. With dim equal to 2, a variable is removed when it holds a missing value, and the options below count down each variable instead of across each row.

tbl = rmmissing (…, Name, Value) customizes the operation with the following options:

'MinNumMissing'

A positive integer n (default 1). A row is removed only when it has at least n variables with a missing value, and with dim equal to 2 a variable only when it is missing in at least n rows.

'DataVariables'

Restrict the search for missing values to the indicated subset of table variables, using the same variable referencing as the other table methods. Variables outside the subset are not inspected, so they are kept in the output whatever they hold.

'MissingLocations'

Supply the missing-value locations explicitly instead of deriving them with ismissing; the values the table holds are then not consulted. The value is either a logical matrix with one row per row of the input and one column per variable of tblA or per inspected variable, or a table of logical variables whose names and sizes match the inspected variables. Where both sizes agree, the columns are taken in the order of the variables of tblA.

[tbl, TF] = rmmissing (…) also returns a logical vector TF that is true for each removed row, a column with one element per row of tblA, or with dim equal to 2 for each removed variable, a row with one element per variable of tblA.

Source Code: table