Access Data in Tables
A table is a container that stores column-oriented data in variables. Table variables can have different data types and sizes as long as all variables have the same number of rows. Table variables have names, just as the fields of a structure have names. The rows of a table can have names, but row names are not required. To access table data, index into the rows and variables using either their names or numeric indices.
Typical reasons for indexing into tables include:
Reordering or removing rows and variables.
Adding arrays as new rows or variables.
Extracting arrays of data to use as input arguments to functions.
Summary of Table Indexing Syntaxes
Depending on the type of indexing you use, you can access either a subtable or an array extracted from the table. Indexing with:
Smooth parentheses, (), returns a table that has selected rows and variables.
Dot notation returns the contents of a variable as an array.
Curly braces, {}, returns an array concatenated from the contents of selected rows and variables.
You can specify rows and variables by name, numeric index, or data type. Starting in
R2019b, variable names and row names can include any characters, including spaces and
non-ASCII characters. Also, they can start with any characters, not just letters. Variable
and row names do not have to be valid MATLAB® identifiers (as determined by the isvarname
function).
Type of Output | Syntax | Rows | Variables | Examples |
|---|---|---|---|---|
Table, containing specified rows and variables |
| Specified as:
| Specified as:
|
|
Table, containing variables that have specified data type |
| Specified as:
| Specified as a data type, such as |
|
Array, extracting data from one variable |
| Not specified | Specified as:
|
|
Array, extracting data from one variable and specified rows |
| Specified as numeric or logical indices of the array | Specified as:
|
|
Array, concatenating data from specified rows and variables |
| Specified as:
| Specified as:
|
|
Array, concatenating data from specified rows and variables with specified data type |
| Specified as:
| Specified as a data type, such as |
|
Array, concatenating data from all rows and variables |
| Not specified | Not specified |
|
Tables Containing Specified Rows and Variables
Load sample data for 100 patients from the patients MAT-file to workspace variables.
load patients
whosName Size Bytes Class Attributes Age 100x1 800 double Diastolic 100x1 800 double Gender 100x1 11412 cell Height 100x1 800 double LastName 100x1 11616 cell Location 100x1 14208 cell SelfAssessedHealthStatus 100x1 11540 cell Smoker 100x1 100 logical Systolic 100x1 800 double Weight 100x1 800 double
Create a table and populate it with the Age, Gender, Height, Weight, and Smoker workspace variables. Use the unique identifiers in LastName as row names. T is a 100-by-5 table. (When you specify row names, they do not count as a table variable).
T = table(Age,Gender,Height,Weight,Smoker,... 'RowNames',LastName)
T=100×5 table
Age Gender Height Weight Smoker
___ __________ ______ ______ ______
Smith 38 {'Male' } 71 176 true
Johnson 43 {'Male' } 69 163 false
Williams 38 {'Female'} 64 131 false
Jones 40 {'Female'} 67 133 false
Brown 49 {'Female'} 64 119 false
Davis 46 {'Female'} 68 142 false
Miller 33 {'Female'} 64 142 true
Wilson 40 {'Male' } 68 180 false
Moore 28 {'Male' } 68 183 false
Taylor 31 {'Female'} 66 132 false
Anderson 45 {'Female'} 68 128 false
Thomas 42 {'Female'} 66 137 false
Jackson 25 {'Male' } 71 174 false
White 39 {'Male' } 72 202 true
Harris 36 {'Female'} 65 129 false
Martin 48 {'Male' } 71 181 true
⋮
Index Using Numeric Indices
Create a subtable containing the first five rows and all the variables from T. To specify the desired rows and variables, use numeric indices within parentheses. This type of indexing is similar to indexing into numeric arrays.
T1 = T(1:5,:)
T1=5×5 table
Age Gender Height Weight Smoker
___ __________ ______ ______ ______
Smith 38 {'Male' } 71 176 true
Johnson 43 {'Male' } 69 163 false
Williams 38 {'Female'} 64 131 false
Jones 40 {'Female'} 67 133 false
Brown 49 {'Female'} 64 119 false
T1 is a 5-by-5 table.
In addition to numeric indices, you can use row or variable names inside the parentheses. (In this case, using row indices and a colon is more compact than using row or variable names.)
Index Using Names
Select all the data for the patients with the last names 'Williams' and 'Brown'. Since T has row names that are the last names of patients, index into T using row names.
T2 = T({'Williams','Brown'},:)T2=2×5 table
Age Gender Height Weight Smoker
___ __________ ______ ______ ______
Williams 38 {'Female'} 64 131 false
Brown 49 {'Female'} 64 119 false
T2 is a 2-by-5 table.
You also can select variables by name. Create a table that has only the first five rows of T and the Height and Weight variables. Display it.
T3 = T(1:5,{'Height','Weight'})T3=5×2 table
Height Weight
______ ______
Smith 71 176
Johnson 69 163
Williams 64 131
Jones 67 133
Brown 64 119
Table variable names do not have to be valid MATLAB identifiers. They can include spaces and non-ASCII characters, and can start with any character.
Add a variable name with spaces and a dash to T. Then index into T using variable names.
T = addvars(T,SelfAssessedHealthStatus,'NewVariableNames','Self-Assessed Health Status'); T(1:5,{'Age','Smoker','Self-Assessed Health Status'})
ans=5×3 table
Age Smoker Self-Assessed Health Status
___ ______ ___________________________
Smith 38 true {'Excellent'}
Johnson 43 false {'Fair' }
Williams 38 false {'Good' }
Jones 40 false {'Fair' }
Brown 49 false {'Good' }
Specify Data Type Subscript
Instead of specifying variables using names or numbers, you can create a data type subscript that matches all variables having the same data type.
First, create a data type subscript to match numeric table variables.
S = vartype('numeric')S = table vartype subscript: Select table variables matching the type 'numeric' See Access Data in a Table.
Create a table that has only the numeric variables, and only the first five rows, from T.
T4 = T(1:5,S)
T4=5×3 table
Age Height Weight
___ ______ ______
Smith 38 71 176
Johnson 43 69 163
Williams 38 64 131
Jones 40 67 133
Brown 49 64 119
Extract Data Using Dot Notation and Logical Values
Create a table from the patients MAT-file. Then use dot notation to extract data from table variables. You can also index using logical indices generated from values in a table variable that meet a condition.
load patients T = table(Age,Gender,Height,Weight,Smoker,... 'RowNames',LastName);
Extract Data from Variable
To extract data from one variable, use dot notation. Extract numeric values from the variable Weight. Then plot a histogram of those values.
histogram(T.Weight)
title('Patient Weight')
T.Weight is a double-precision column vector with 100 rows.
Select Rows with Logical Indexing
You can index into an array or a table using an array of logical indices. Typically, you use a logical expression that determines which values in a table variable meet a condition. The result of the expression is an array of logical indices.
For example, create logical indices matching patients whose age is less than 40.
rows = T.Age < 40
rows = 100x1 logical array
1
0
1
0
0
0
1
0
1
1
⋮
To extract heights for patients whose age is less than 40, index into the Height variable using rows. There are 56 patients younger than 40.
T.Height(rows)
ans = 56×1
71
64
64
68
66
71
72
65
69
69
⋮
You can index into a table with logical indices. Display the rows of T for the patients who are younger than 40.
T(rows,:)
ans=56×5 table
Age Gender Height Weight Smoker
___ __________ ______ ______ ______
Smith 38 {'Male' } 71 176 true
Williams 38 {'Female'} 64 131 false
Miller 33 {'Female'} 64 142 true
Moore 28 {'Male' } 68 183 false
Taylor 31 {'Female'} 66 132 false
Jackson 25 {'Male' } 71 174 false
White 39 {'Male' } 72 202 true
Harris 36 {'Female'} 65 129 false
Thompson 32 {'Male' } 69 191 true
Garcia 27 {'Female'} 69 131 true
Martinez 37 {'Male' } 70 179 false
Rodriguez 39 {'Female'} 64 117 false
Walker 28 {'Female'} 65 123 true
Hall 25 {'Male' } 70 189 false
Allen 39 {'Female'} 63 143 false
Young 25 {'Female'} 63 114 false
⋮
You can match multiple conditions with one logical expression. Display the rows for smoking patients younger than 40.
rows = (T.Smoker==true & T.Age<40); T(rows,:)
ans=18×5 table
Age Gender Height Weight Smoker
___ __________ ______ ______ ______
Smith 38 {'Male' } 71 176 true
Miller 33 {'Female'} 64 142 true
White 39 {'Male' } 72 202 true
Thompson 32 {'Male' } 69 191 true
Garcia 27 {'Female'} 69 131 true
Walker 28 {'Female'} 65 123 true
King 30 {'Male' } 67 186 true
Nelson 33 {'Male' } 66 180 true
Mitchell 39 {'Male' } 71 164 true
Turner 37 {'Male' } 70 194 true
Sanders 33 {'Female'} 67 115 true
Price 31 {'Male' } 72 178 true
Jenkins 28 {'Male' } 69 189 true
Long 39 {'Male' } 68 182 true
Patterson 37 {'Female'} 65 120 true
Flores 31 {'Female'} 66 141 true
⋮
Dot Notation with Any Variable Name or Expression
When you index using dot notation, there are two ways to specify a variable.
By name, without quotation marks. For example,
T.Datespecifies a variable named'Date'.By an expression, where the expression is enclosed by parentheses after the dot. For example,
T.('Start Date')specifies a variable named'Start Date'.
Use the first syntax when a table variable name also happens to be a valid MATLAB® identifier. (A valid identifier starts with a letter and includes only letters, digits, and underscores.)
Use the second syntax when you specify:
A number that indicates the position of the variable in the table.
A variable name that isn't a valid MATLAB identifier.
A function whose output is the name of a variable in the table, or a variable you add to the table. The output of the function must be a character vector or a string scalar.
For example, create a table from the patients MAT-file. Then use dot notation to access the contents of table variables.
load patients T = table(Age,Gender,Height,Weight,Smoker,... 'RowNames',LastName);
To specify a variable by position in the table, use a number. Age is the first variable in T, so use the number 1 to specify its position.
T.(1)
ans = 100×1
38
43
38
40
49
46
33
40
28
31
⋮
To specify a variable by name, you can enclose it in quotation marks. Since 'Age' is a valid identifier, you can specify it using either T.Age or T.('Age').
T.('Age')ans = 100×1
38
43
38
40
49
46
33
40
28
31
⋮
You can specify table variable names that are not valid MATLAB identifiers. Variable names can include spaces and non-ASCII characters, and can start with any character. However, when you use dot notation to access a table variable with such a name, you must specify it using parentheses.
Add a variable name with spaces and a hyphen to T.
T = addvars(T,SelfAssessedHealthStatus,'NewVariableNames','Self-Assessed Health Status'); T(1:5,:)
ans=5×6 table
Age Gender Height Weight Smoker Self-Assessed Health Status
___ __________ ______ ______ ______ ___________________________
Smith 38 {'Male' } 71 176 true {'Excellent'}
Johnson 43 {'Male' } 69 163 false {'Fair' }
Williams 38 {'Female'} 64 131 false {'Good' }
Jones 40 {'Female'} 67 133 false {'Fair' }
Brown 49 {'Female'} 64 119 false {'Good' }
Access the new table variable using dot notation. Display the first five elements.
C = T.('Self-Assessed Health Status');
C(1:5)ans = 5x1 cell
{'Excellent'}
{'Fair' }
{'Good' }
{'Fair' }
{'Good' }
You also can use the output of a function as a variable name. Delete the T.('Self-Assessed Health Status') variable. Then replace it with a variable whose name includes today's date.
T.('Self-Assessed Health Status') = []; T.(string(datetime('today')) + ' Self Report') = SelfAssessedHealthStatus; T(1:5,:)
ans=5×6 table
Age Gender Height Weight Smoker 17-Aug-2020 Self Report
___ __________ ______ ______ ______ _______________________
Smith 38 {'Male' } 71 176 true {'Excellent'}
Johnson 43 {'Male' } 69 163 false {'Fair' }
Williams 38 {'Female'} 64 131 false {'Good' }
Jones 40 {'Female'} 67 133 false {'Fair' }
Brown 49 {'Female'} 64 119 false {'Good' }
Extract Data from Specified Rows and Variables
Indexing with curly braces extracts data from a table and results in an array, not a subtable. But other than that difference, you can specify rows and variables using numbers, names, and data type subscripts, just as you can when you index using smooth parentheses. To extract values from a table, use curly braces. If you extract values from multiple table variables, then the variables must have data types that allow them to be concatenated together.
Specify Rows and Variables
Create a table from numeric and logical arrays from the patients file.
load patients T = table(Age,Height,Weight,Smoker,... 'RowNames',LastName);
Extract data from multiple variables in T. Unlike dot notation, indexing with curly braces can extract values from multiple table variables and concatenate them into one array.
Extract the height and weight for the first five patients. Use numeric indices to select the first five rows, and variable names to select the variables Height and Weight.
A = T{1:5,{'Height','Weight'}}A = 5×2
71 176
69 163
64 131
67 133
64 119
A is a 5-by-2 numeric array, not a table.
If you specify one variable name, then curly brace indexing results in the same array you can get with dot notation. However, you must specify both rows and variables when you use curly brace indexing. For example, this syntaxes T.Height and T{:,'Height'} return the same array.
Extract Data from All Rows and Variables
If all the table variables have data types that allow them to be concatenated together, then you can use the T.Variables syntax to put all the table data into an array. This syntax is equivalent to T{:,:} where the colons indicate all rows and all variables.
A2 = T.Variables
A2 = 100×4
38 71 176 1
43 69 163 0
38 64 131 0
40 67 133 0
49 64 119 0
46 68 142 0
33 64 142 1
40 68 180 0
28 68 183 0
31 66 132 0
⋮
See Also
addvars | histogram | table | vartype