Using Stata Automation

You can write Microsoft Windows applications to interact with Stata via Automation.

The following topics are discussed.

1. What is Stata Automation?

Automation (formerly known as OLE Automation) is a communication mechanism between Microsoft Windows applications. It provides an infrastructure whereby Windows applications (Automation clients) can access and manipulate functions and properties implemented in another application (Automation server). A Stata Automation object exposes internal Stata methods and properties so that Windows programmers can write Automation clients to directly use the services provided by Stata.

2. When is Stata Automation beneficial?

A Stata Automation object is most useful for situations that require the greatest flexibility to interact with Stata from user-written Windows applications. A Stata Automation object enables users to directly access Stata macros, scalars, saved results, and datasets in ways besides the usual log files.

3. Create a Microsoft Windows application to interact with Stata via Automation

3.1 Install Stata Automation object

To use a Stata Automation object, the type library has to be registered after Stata is installed on a Microsoft Windows machine.

To register the Stata Automation type library on a Windows non-Vista machine, do the following steps:

1. From the Start menu, select Run, enter cmd, and click OK.
2. Change the directory to the directory with the Stata executable.
3. Type StataSE /Register or StataSE /Regserver.

To register the Stata Automation type library on Windows Vista, do the following steps:

1. In the installation directory, right-click on the Stata executable, for example, StataSE.exe. Choose "Create Shortcut".
2. Right click on the newly created "Shortcut to StataSE.exe", choose "Property", and change the Target from "C:\Program Files\Stata12\StataSE.exe" to "C:\Program Files\Stata12\StataSE.exe" /Register. Click "OK".
3. Right-click on the updated "Shortcut to StataSE.exe"; choose "Run as administrator".

The following steps will unregister the Stata Automation type library:

1. From the Start menu, select Run, enter cmd, and click OK.
2. Change the directory to the directory with the Stata executable.
3. Type StataSE /Unregister or StataSE /Unregserver.

To manually register the type library, do the following steps:

1. From the Start menu, select Run, enter cmd, and click OK.
2. Change the directory to the directory with the Stata executable.
3. Type StataSE /Register or StataSE /Regserver.

3.2 Load Stata Automation object into your project via early binding

For VB.NET or C#:

1. In Visual Studio Solution Explorer, select the project.
2. On the Project menu, choose Add Reference. The Add Reference dialog box opens.
3. Select the COM tab.
4. In the top pane, select the component name stata.

See http://msdn2.microsoft.com/en-us/library/7314433t(VS.80).aspx for more information.

5. In the application, the user may create a Stata Automation object as follows:

           [VB.NET]
           dim stataobj As New stata.StataOLEApp()	

3.3 Load Stata Automation object via late binding

For VBScript, the user may use the VBScript method CreateObject() to load a Stata Automation object as follows:

        [VBScript]
        dim stataobj
        set stataobj=CreateObject("stata.StataOLEApp")

Note: Stata Automation is a single-use out-of-process server. This means that creating a new Automation object will launch a new instance of Stata. It is strongly recommended that your software create one Stata Automation object and use it through the execution.

3.4 Control Stata’s display mode

When a Stata Automation object is created, the Stata main window is displayed. Method UtilShowStata() could be used to hide or minimize the Stata main window.

        [VB.NET]
        ' create a Stata Automation object and Stata is launched 
        ' and will be visible. 
        Dim stataobj As New stata.StataOLEApp()
        ' hide Stata
        stataobj.UtilShowStata(1)
        MsgBox("Stata is hidden")
        ' minimized Stata
        stataobj.UtilShowStata(2)
        MsgBox("Stata is minimized")
        ' restored Stata
        stataobj.UtilShowStata(0)
        MsgBox("Stata is restored")
        stataobj = Nothing

3.5 Release Stata Automation objects

The programmer must release the Stata Automation object to avoid memory leaks.

        [VB.NET]
        stataobj = nothing

        [VBScript]
        set stataobj = nothing	

3.6 Run Stata commands

The programmer may submit a Stata command from the client to Stata via Automation in two ways. One uses DoCommand(), and the other uses DoCommandAsync(). DoCommand() returns the return code after the command is finished in Stata and blocks the client during execution. DoCommandAsync() executes a command asynchronously.

DoCommand() returns the Stata return code from executing the Stata command. A nonzero value means that execution of the command failed.

DoCommand() ignores Stata command exit. The method returns error code 10001 for Stata command program define, while, forvalues, foreach, and input. The programmer who uses Stata Automation needs to pay attention to Stata methods and program that might block Stata for further user input. Because the Stata window may be hidden or minimized, the end user will not be able to type commands into the Stata Command window. Note: program define, while, forvalues, foreach, and input can still be used inside do- or ado-files.

        [VB.NET]
        dim rc as Integer
        dim stataobj as new stata.StataOLEApp()
        rc = stataobj.DoCommand("sysuse auto, clear")
        if rc = 0 then
                rc = stataobj.DoCommand("regress mpg price")
        else
                ' do error handling
        end if
        stataobj = nothing

3.7 Run Stata commands asynchronously

DoCommandAsync() executes a command asynchronously. It copies the Stata command into a queue and then returns immediately. The command will be executed by Stata when appropriate. The method is equivalent to typing in the Stata Command window. The method has a major advantage over DoCommand(): the user may submit program define, while, forvalues, foreach, and input.

Because the method executes the Stata command asynchronously, the user must be careful about mixing it with DoCommand() and synchronizing the client and server. Sometimes the programmer must make sure that the command submitted via DoCommandAsync() has finished before proceeding to the next operation.

        [VB.NET]
        dim stataobj as new stata.StataOLEApp()
        ' program define will fail in DoCommand()
        stataobj.DoCommandAsync("program define myreg")
        stataobj.DoCommandAsync("regress mpg price")
        stataobj.DoCommandAsync("end")
        stataobj.DoCommandAsync("sysuse auto, clear")
        stataobj.DoCommandAsync("myreg")
        stataobj = nothing

Here we give an example of determining whether Stata has finished executing a command. The method to use is UtilIsStataFreeEvent(). It monitors the status of Stata and sends an OnFinish event to the client when Stata finishes the execution.

        [VB.NET]
        ' to use WithEvents, stataobj cannot have local scope
        Public WithEvents stataobj As stata.StataOLEApp
        Public IsStataFree As Boolean

        Private Sub testCommandAsync()
                Dim eN As Integer
                Dim obs As Integer
                Dim var As Integer

                ' create Stata Automation object
                stataobj = New stata.StataOLEApp()

                IsStataFree = False
                obs = 50000
                var = 100
                stataobj.DoCommandAsync("clear")
                stataobj.DoCommandAsync("set more off")
                stataobj.DoCommandAsync("set seed 12345")
                stataobj.DoCommandAsync("set mem 400M")

                stataobj.DoCommandAsync("local obs " & obs)
                stataobj.DoCommandAsync("local var " & var)
                stataobj.DoCommandAsync("set obs `obs'")

                stataobj.DoCommandAsync("gen y = 0")
                stataobj.DoCommandAsync("qui forval i=1/`var' {")
                stataobj.DoCommandAsync("gen x`i' = runiform()")
                stataobj.DoCommandAsync("replace y= y+`i'*x`i'")
                stataobj.DoCommandAsync("}")

                stataobj.DoCommandAsync("regress y x*")
                ' make Stata monitor its own status and fire an OnFinish event
                ' when above commands are finished. The OnFinish event is 
                ' listened and handled by stataobj_OnFinish().
                stataobj.UtilIsStataFreeEvent()
                While IsStataFree = False
                        Application.DoEvents()
                End While
        	
                ' Since the value of e(N) depends on the execution of -regress-,
                ' it is very important to make sure the -regress- has finished.
                ' Otherwise, eN will not get the correct value. 
                eN = stataobj.StReturnNumeric("e(N)")
                MsgBox("using DoCommandAsync e(N) = " & eN)

                stataobj = Nothing
        End Sub

        Private Sub Button1_Click(ByVal sender As System.Object, ByVal e As
        System.EventArgs) Handles Button1.Click
                testCommandAsync()
        End Sub

        ' handles OnFinish event.
        Private Sub stataobj_OnFinish(ByVal noCommand As Integer) Handles
        stataobj.OnFinish
                IsStataFree = True
        End Sub

3.8 Make Stata stop what it is doing

It is useful to have a way to stop Stata when it is executing a long command. There are two ways to do so.

If Stata is visible, the user can stop the execution by clicking the Break button (the button with the big red X) or by holding down Ctrl and pressing the Pause/Break key.

The method UtilSetStataBreak() can also break the execution. Using UtilSetStataBreak() with DoCommand() can be tricky since DoCommand() blocks the client. Hence the client needs a worker thread to run the DoCommand() and uses UtilSetStataBreak() from the main thread. On the other hand, using UtilSetStataBreak() with DoCommandAsync() is easy.

Here is an example. The DoCommandAsync() sends a long command to Stata. The user could cancel the execution by clicking the cancel button.

        [VB.NET] 
        ' the lengthy command sent to Stata is to execute the following do-file,
        ' test.do	 
        '       // test.do
        '       clear
        '       set more off
        '       set seed 12345
        '       set mem 200M
        '       local obs 50000
        '       local var 100
        '       set obs `obs'
        '       gen y=0
        '       qui forval i=1/`var' {
        '		gen x`i' = runiform()
        '		replace y= y+`i'*x`i'
        '       }
        '       regress y x*
        '
        '       //end of test.do
        '
        ' to use WithEvents, stataobj cannot have local scope
        Public WithEvents stataobj As stata.StataOLEApp
        Public IsStataFree As Boolean

        Private Sub testCommandAsync()
                ' create Stata Automation object
                stataobj = New stata.StataOLEApp()

                stataobj.DoCommandAsync("do test.do")
                ' make Stata monitor its own status and fire an OnFinish event
                ' when above commands are finished. 
                stataobj.UtilIsStataFreeEvent()
                While IsStataFree = False
                        Application.DoEvents()
                End While
                stataobj = Nothing
        End Sub

        Private Sub butCancel_Click(ByVal sender As System.Object, ByVal e As
        System.EventArgs) Handles butCancel.Click
                If stataobj Is Nothing Then
                        ' do nothing
                Else
                        stataobj.UtilSetStataBreak()
                End If
        End Sub

3.9 Get macros, scalars, and saved results from Stata

3.9.1 Get macros from Stata

Method MacroValue() gets the value of a macro from Stata. The macro value is returned as a string. If the macro with the specified name cannot be found, the method returns empty string.

The name of local macro needs to be prefixed with “_” when submitted via MacroValue() because Stata automatically prefixes local macros with “_”.

        [VB.NET]
        Imports System

        ' get macro 
        Dim stataobj As New stata.StataOLEApp()
        Dim mGlobal, mLocal As String
        stataobj.DoCommand("global mGlobal = ""test global macro""")
        stataobj.DoCommand("local mLocal = ""test local macro""")

        ' get global macro mGlobal
        mGlobal = stataobj.MacroValue("mGlobal")
        ' get local macro mLocal. Note: local macro name need have a "_" prefix.
        mLocal = stataobj.MacroValue("_mLocal")
        Console.WriteLine("global macro mGlobal : " & mGlobal & ";" & "local 
        > macro mLocal : " & mLocal & vbCrLf)
        ' if the macro name is wrong, MacroValue returns empty string.
        mLocal = stataobj.MacroValue("mLocal")
        Console.WriteLine("local macro mLocal with wrong name(returns empty
        > string): " & mLocal)
        stataobj = Nothing

3.9.2 Get scalars from Stata

Methods ScalarNumeric() and ScalarString() get the value of a scalar from Stata. Method ScalarType() determines whether the scalar is numeric or string.

        [VB.NET]
        Imports System
        ' get scalar
        Dim stataobj As New stata.StataOLEApp()
        stataobj.DoCommand("scalar na=10")
        stataobj.DoCommand("scalar sa=""this is a test""")
        Dim na As Double
        Dim sa As String
        Dim type As Integer
        '
        ' type = 0 : the scalar with the name is not found.
        '        1 : the scalar with the name is numeric.
        '        2 : the scalar with the name is string.
        '
        type = stataobj.ScalarType("na")
        If type = 1 Then
            Console.WriteLine("scalar na is of numeric type")
        End If
        na = stataobj.ScalarNumeric("na")    ' na is 10 now
        sa = stataobj.ScalarString("sa")     ' sa is "this is a test" now 
        Console.WriteLine("numeric na=" & na & "; string sa=" & sa)
        ' get the scalar with wrong data type
        sa = stataobj.ScalarNumeric("sa")    ' sa is Stata missing value now
        na = stataobj.ScalarString("na")     ' na is 0 now, since it has the 
                                             ' double value casted from an 
                                             ' empty string.
        
        Console.WriteLine("wrong type:numeric na=" & na & ";string sa=" & sa)
        ' test for nonexistent scalars 
        na = stataobj.ScalarNumeric("not_exist")    ' na is misval
        sa = stataobj.ScalarString("not_exist")     ' sa is empty string
        Console.WriteLine("not exist: numeric na=" & na & ";string sa=" & sa)
        stataobj = Nothing

3.9.3 Get saved results from Stata

Methods StReturnNumeric() and StReturnString() get the value of a saved result from Stata. Method StReturnType() determines whether the saved result is numeric or string. Section 3.11 discusses the saved result matrix.

        [VB.NET]
        Imports System
        Dim stataobj As New stata.StataOLEApp()
        ' get saved return and type

        ' Types of return:
        '   15 - numeric
        '   3059 - string

        Dim c_current_date As String
        Dim c_current_date_type As Integer
        Dim c_max_matsize As Integer
        Dim c_max_matsize_type As Integer
        Dim c_maxdouble As Double
        Dim c_maxdouble_type As Integer

        c_current_date = stataobj.StReturnString("c(current_date)")
        c_current_date_type = stataobj.StReturnType("c(current_date)")
        c_max_matsize = stataobj.StReturnNumeric("c(max_matsize)")
        c_max_matsize_type = stataobj.StReturnType("c(max_matsize)")
        c_maxdouble = stataobj.StReturnNumeric("c(maxdouble)")
        c_maxdouble_type = stataobj.StReturnType("c(maxdouble)")

        Console.WriteLine("current_date type:" &c_current_date_type & ";" )
        Console.WriteLine("current_date:" & c_current_date & vbCrLf)
        Console.WriteLine("max_matsize type:" &c_max_matsize_type & ";" )
        Console.WriteLine("max_matsize:" & c_max_matsize & vbCrLf)
        Console.WriteLine("maxdouble type:" & c_maxdouble_type & ";")
        Console.WriteLine("max_double:" & c_maxdouble & vbCrLf)

        ' get e return
        Dim e_N As Integer
        Dim e_N_type As Integer
        Dim e_F As Double
        Dim e_F_type As Integer
        Dim e_depvar As String
        Dim e_depvar_type As Integer
        stataobj.DoCommand("sysuse auto, clear")
        stataobj.DoCommand("regress mpg price")
        e_N = stataobj.StReturnNumeric("e(N)")
        e_N_type = stataobj.StReturnType("e(N)")
        e_F = stataobj.StReturnNumeric("e(F)")
        e_F_type = stataobj.StReturnType("e(F)")
        e_depvar = stataobj.StReturnString("e(depvar)")
        e_depvar_type = stataobj.StReturnType("e(depvar)")

        Console.WriteLine("e(N) type:" & e_N_type & ";")
        Console.WriteLine("e(N):" & e_N & vbCrLf)
        Console.WriteLine("e(F) type:" & e_F_type & ";")
        Console.WriteLine("e(F):" & e_F & vbCrLf)
        Console.WriteLine("e(depvar) type:" & e_depvar_type & ";")
        Console.WriteLine("e(depvar):" & e_depvar & vbCrLf)

        ' get r return
        Dim r_N As Integer
        Dim r_N_type As Integer
        Dim r_mean As Double
        Dim r_mean_type As Integer
        stataobj.DoCommand("sysuse auto, clear")
        stataobj.DoCommand("summarize")
        r_N = stataobj.StReturnNumeric("r(N)")
        r_N_type = stataobj.StReturnType("r(N)")
        r_mean = stataobj.StReturnNumeric("r(mean)")
        r_mean_type = stataobj.StReturnType("r(mean)")

        Console.WriteLine("r(N) type:" & r_N_type & ";")
        Console.WriteLine("r(N):" & r_N & vbCrLf)
        Console.WriteLine("r(mean) type:" & r_mean_type & ";")
        Console.WriteLine("r(mean):" & r_mean & vbCrLf)

        stataobj = Nothing

3.10 Handle Stata missing value

Stata system missing value “.” is represented by an arbitrarily large positive value. Any number greater than or equal to this value will be treated as a missing value in Stata. There are two ways to determine whether a number will be treated as a missing value.

Method UtilIsStMissingValue(double valueToCheck) returns True if the valueToCheck is greater than or equal to the Stata system missing value.

Programmers can also use method UtilGetStMissingValue() to get a local copy of the Stata system missing value and then compare numbers with it in the client side. This way saves future round trips of sending data to the Stata server.

        [VB.NET]
        Imports System
        Dim stataobj As New stata.StataOLEApp()
        Dim misVal, misVal1, misVal2 As Double
        Dim isMissing As Boolean

        misVal = stataobj.UtilGetStMissingValue()
        isMissing = stataobj.UtilIsStMissingValue(misVal)
        If isMissing = True Then        ' is True
            Console.WriteLine(misVal & " is treated as missing." & vbCrLf)
        End If

        misVal1 = 5.0E+307
        isMissing = stataobj.UtilIsStMissingValue(misVal1)
        If isMissing = False Then       ' is false
            Console.WriteLine(misVal1 & " is not treated as missing." & vbCrLf)
        End If

        misVal2 = 2 * misVal1
        isMissing = stataobj.UtilIsStMissingValue(misVal2)
        If isMissing = True Then        ' is True
            Console.WriteLine(misVal2 & " is treated as missing." & vbCrLf)
        End If
        stataobj = Nothing

3.11 Get matrix from Stata

Method MatrixData() returns a Stata matrix as a one-dimensional double array. The size of the array is the number of entries in the matrix. The method returns the full matrix even for a symmetric matrix.

Methods MatrixRowDim() and MatrixColDim() return the number of rows and the number of columns of the matrix, respectively. Methods MatrixRowNames() and MatrixColNames() return the row names and the column names of the matrix in a string array.

        [VB.NET]
        Imports System
        ' get matrix
        Dim stataobj As New stata.StataOLEApp()
        Dim eV As Double()
        Dim i, j, t As Integer
        Dim row, column As Integer
        Dim rowNames, columnNames As String()

        stataobj.DoCommand("sysuse auto, clear")
        stataobj.DoCommand("regress mpg price")
        ' get number of rows of e(V)
        row = stataobj.MatrixRowDim("e(V)")
        ' get row names of matrix e(V)
        rowNames = stataobj.MatrixRowNames("e(V)")
        ' get number of columns of e(V)
        column = stataobj.MatrixColDim("e(V)")
        ' get column names of matrix e(V)
        columnNames = stataobj.MatrixColNames("e(V)")
        ' get matrix
        eV = stataobj.MatrixData("e(V)")

        ' display matrix 
        Console.WriteLine("matrix e(V)[" & row & ", " & column & "]" & vbCrLf)
        ' display column header
        Console.WriteLine(String.Format("{0,20}", ""))
        For i = 0 To column - 1
            Console.WriteLine(String.Format("{0,20}", columnNames(i)))
        Next
        Console.WriteLine(vbCrLf)
        For i = 0 To row - 1
            ' display row header
            Console.WriteLine(String.Format("{0,-20}", rowNames(i)))
            t = i * column
            For j = 0 To column - 1
                Console.WriteLine(String.Format("{0,20}", eV(t).ToString("e8")))
                t = t + 1
            Next
            Console.WriteLine(vbCrLf)
        Next
        MsgBox("Got matrix information for e(V)")
        stataobj = Nothing

3.12 Get data information from Stata

3.12.1 Get variable index and name

Method StVariableIndex() returns the index of the variable given the name. If the variable with the name is not found, StVariableIndex() returns −111.

Method StVariableName() returns the name of the variable given the index. A valid variable index is greater than or equal to 1 and is less than or equal to the number of variables. If the variable index is not valid, StVariableName() returns an empty string. Method VariableNameArray() returns a string array of all variable names.

        [VB.NET]
        Imports System
        ' get variable index
        Dim stataobj As New stata.StataOLEApp()
        Dim varIndex As Integer

        stataobj.DoCommand("sysuse auto, clear")
        ' get the variable index of make
        varIndex = stataobj.StVariableIndex("make")
        Console.WriteLine("make-variable index:" & varIndex & vbCrLf)
        ' if the variable cannot be found
        varIndex = stataobj.StVariableIndex("not_exist")
        Console.WriteLine("not_exist-variable index:" & varIndex & vbCrLf)

        ' get variable name
        Dim varName As String
        ' get the first variable name
        varName = stataobj.StVariableName(1)
        Console.WriteLine("var 1 name:" & varName & vbCrLf)
        ' get the last variable name
        varName = stataobj.StVariableName(12)
        Console.WriteLine("var 12 name:" & varName & vbCrLf)
        ' try an out-of-range variable index 
        varName = stataobj.VariableLabel("13")
        Console.WriteLine("var 13 (out of range) name:" & varName & vbCrLf)

        ' get all variable names in an array
        Dim varNames As String()
        Dim i As Integer
        ' get all variable names
        varNames = stataobj.VariableNameArray()
        For i = 0 To varNames.Length - 1
            Console.WriteLine("var " & i + 1 & " name:" & varNames(i) & vbCrLf)
        Next
        stataobj = Nothing

3.12.2 Get variable data type

Method VariableType() returns an integer that represents the data type of the variable. Possible data type values include the following:

        0:      type not defined
        1-244:  string variable with maximum length of the value
        251:    byte
        252:    integer
        253:    long integer
        254:    float
        255:    double

The result can be compared with the values in stata.StataVariableTypes enum type.

Method VariableTypeArray() returns a data type array for all variables.

        [VB.NET]
        Imports System
        ' get variable data type
        Dim stataobj As New stata.StataOLEApp()
        Dim varType As Byte

        stataobj.DoCommand("sysuse auto, clear")
        ' get the type of make
        varType = stataobj.VariableType("make")
        ' vartype for make is 18. Test against stata.StataVariableTypes   
        If varType = stata.StataVariableTypes.vstUndef Then
            Console.WriteLine("data type undefined")
        End If

        If varType <= stata.StataVariableTypes.vstString And varType >= 1 Then
            Console.WriteLine("data type string, maximum length " & varType)
        End If

        If varType = stata.StataVariableTypes.vstByte Then
            Console.WriteLine("data type byte")
        End If

        If varType = stata.StataVariableTypes.vstInt Then
            Console.WriteLine("data type integer")
        End If

        If varType = stata.StataVariableTypes.vstLong Then
            Console.WriteLine("data type long")
        End If

        If varType = stata.StataVariableTypes.vstFloat Then
            Console.WriteLine("data type float")
        End If

        If varType = stata.StataVariableTypes.vstDouble Then
            Console.WriteLine("data type double")
        End If
        Console.WriteLine(vbCrLf)
        
        ' get the type of price
        varType = stataobj.VariableType("price")
        ' vartype for price is 252, which corresponds to integer.
        Console.WriteLine("var price-data type:" & varType & vbCrLf)
        ' vartype is 0 if it cannot be found.
        varType = stataobj.VariableType("not_exist")
        Console.WriteLine("var not_exist-data type:" & varType & vbCrLf)

        ' get data type array for all variables
        Dim varTypes As Byte()
        Dim i As Integer

        ' get the types of all variables
        varTypes = stataobj.VariableTypeArray()
        For i = 0 To varTypes.Length - 1
            Console.WriteLine("var " &i+1&"-data type:" & varTypes(i) & vbCrLf)
        Next
        stataobj = Nothing

3.12.3 Get variable label

Method VariableLabel() returns the label of a variable. Method VariableLabelArray() returns labels for all variables.

        [VB.NET]
        Dim stataobj As New stata.StataOLEApp()
        Dim varLabel As String

        stataobj.DoCommand("sysuse auto, clear")
        ' get the variable label of make
        varLabel = stataobj.VariableLabel("make")
        Console.WriteLine("var make-label:" & varLabel & vbCrLf)
        ' test for nonexistent variables
        varLabel = stataobj.VariableLabel("not_exist")
        Console.WriteLine("var not_exist-label:" & varLabel & vbCrLf)

        ' get variable labels for all variables
        Dim varLabels As String()
        Dim i As Integer

        ' get all variable labels
        varLabels = stataobj.VariableLabelArray()
        For i = 0 To varLabels.Length - 1
            Console.WriteLine("var " &i+1& "-label:" & varLabels(i) & vbCrLf)
        Next
        stataobj = Nothing

3.12.4 Get value label

Method VariableValLabel() returns the value label for a variable. It returns the empty string if the variable does not have a value label.

Method VariableValLabelArray() returns a one-dimensional string array for all value labels for all variables.

A value label contains several value–label pairs. Method VariableValLabelCount() returns the number of value–label pairs in a value label. Method VariableValLabelValueArray() returns an integer array of all values in a value label. Method VariableValLabelLabelArray() returns a string array of all labels in a value label.

        [VB.NET]
        Imports System
        ' test value labels
        Dim stataobj As New stata.StataOLEApp()
        Dim valLabelName As String
        Dim varIndex As Integer

        stataobj.DoCommand("sysuse auto, clear")
        ' get the value label of variable foreign
        varIndex = stataobj.StVariableIndex("foreign")
        If varIndex > 0 Then
            ' get the variable value label
            valLabelName = stataobj.VariableValLabel(varIndex)
            ' valLabelName is -origin-
            Console.WriteLine("var foreign-value label:"&valLabelName& vbCrLf)
        End If

        ' get the value label of variable make
        varIndex = stataobj.StVariableIndex("make")
        If varIndex > 0 Then
            ' get the variable value label
            valLabelName = stataobj.VariableValLabel(varIndex)
            ' valLabelName is empty string
            Console.WriteLine("var make-value label:" & valLabelName & vbCrLf)
        End If

        Dim valLabelNames As String()
        Dim i As Integer

        ' get the value labels
        valLabelNames = stataobj.VariableValLabelArray()
        For i = 0 To valLabelNames.Length - 1
            Console.WriteLine("value label-"&i+1& &valLabelNames(i)& vbCrLf)
        Next

        ' value label origin has two values, 0 and 1, and one label for each value. 
        ' get the number of values in a value label
        Dim valCount As Integer
        valCount = stataobj.VariableValLabelCount("origin")
        Console.WriteLine("value label origin has " &valCount& " values"&vbCrLf)

        ' get each label
        Dim valLabel As String
        For i = 0 To valCount - 1
            valLabel = stataobj.VariableValLabelForValue("origin", i)
            Console.WriteLine("origin "&i+1& "-th label is: "&valLabel& vbCrLf)
        Next

        ' get all values in an array and all labels in another array
        Dim values As Integer()
        Dim labels As String()

        ' get all values in value label origin
        values = stataobj.VariableValLabelValueArray("origin")
        ' get all labels in value label origin
        labels = stataobj.VariableValLabelLabelArray("origin")
        For i = 0 To values.Length - 1
            Console.WriteLine("origin " &i+1& "-th value is: "&values(i)&vbCrLf)
            Console.WriteLine("origin " &i+1& "-th label is: "&values(i)&vbCrLf)
        Next
        stataobj = Nothing

3.13 Get data from Stata

3.13.1 Get data of one variable from Stata

Methods VariableData() and VariableDataFromName() return an observation array for a variable. The returned array contains variants. The array can be either string or double. Programmers need to determine the data type and make the correct type cast.

Both methods take the start observation index startObs and end observation index endObs as the second and the third arguments. Let nobs be the number of observations in the dataset. The following relations must hold for inputs:

1. startObs ≥ 1
2. endObs ≤ nobs
3. startObs ≤ endObs

Otherwise, both methods produce an ERROR_INVALID_PARAMETER error and return a NULL array.

        [VB.NET]
        Imports System
        ' get data from one variable
        Dim stataobj As New stata.StataOLEApp()
        Dim varName As String
        Dim varType As Integer
        Dim nobs As Integer
        Dim i As Integer

        stataobj.DoCommand("sysuse auto, clear")
        ' get the type and data of the first variable - make
        varName = stataobj.StVariableName(1)
        varType = stataobj.VariableType(varName)
        ' get number of observations
        stataobj.DoCommand("scalar nobs=_N")
        nobs = stataobj.ScalarNumeric("nobs")
        If varType = stata.StataVariableTypes.vstUndef Then
            ' undefined data type - do error handling
            Exit Sub
        End If

        If varType <= stata.StataVariableTypes.vstString And varType >= 1 Then
            ' string variable
            Dim varData As String()
            ' get all observations             
            varData = stataobj.VariableData(1, 1, nobs)
            For i = 0 To nobs - 1
                Console.WriteLine(i + 1 & ":" & varData(i) & vbCrLf)
            Next
        End If

        ' get the type and data of variable price
        varType = stataobj.VariableType("price")
        ' get number of observations
        stataobj.DoCommand("scalar nobs=_N")
        nobs = stataobj.ScalarNumeric("nobs")
        If varType = stata.StataVariableTypes.vstUndef Then
            ' undefined data type - do error handling
            Exit Sub
        End If

        If varType > stata.StataVariableTypes.vstString Then
            ' numeric variable
            Dim varData As Double()
            ' get all observations
            varData = stataobj.VariableDataFromName("price", 1, nobs)
            For i = 0 To nobs - 1
                Console.WriteLine(i + 1 & ":" & varData(i) & vbCrLf)
            Next
        End If
        stataobj = Nothing

3.13.2 Get data of many variables from Stata

Method VariablesData() returns a variant array of observations of several variables at once. The arguments of the method are

1. startVariableIndex
2. endVariableIndex
3. startObservationIndex
4. endObservationIndex

They must satisfy the following conditions:

1. 1 ≤ startVariableIndex ≤ endObservationIndex ≤ numberofVariables
2. 1 ≤ startObservationIndex ≤ endObservationIndex ≤ numberofObservation

Otherwise, the method produces an INVALID_PARAMETER error and returns a NULL array. The method returns observations for all variables with index from startVariableIndex to endVariableIndex.

Because the variables may be both string and numeric, a variant array must be used instead of a regular two-dimensional array. To get the data for each variable, the programmer must do the following steps:

        ' declare dataArr as a variant array
        dim dataArr() as object 
        dim varType as integer
        dim i as integer
        ' get observations of variable 1 and 2
        dataArr = stataobj.VariablesData(1, 2, 1, 20)	
        dim dataString as String() 
        dim dataDbl as double()
		
        'find the data type of the first variable
        varType = stataobj.VariableType(1)
        ' suppose it is a string variable, assign the first element
        ' dataArr to a string array 
        dataString = dataArr(0)
        ' if it is numeric, assign it to dataDbl	
        ' then you may access each observation of the first variable through 
        ' elements of dataString
        for i=0 to dataString.length-1
                Console.WriteLine(dataString(i))
        next 

Here is an example.

        [VB.NET]
        Imports System
        ' get data of several variables at once
        Dim stataobj As New stata.StataOLEApp()
        Dim varName As String
        Dim varType As Integer
        Dim nobs As Integer
        Dim nvar As Integer
        Dim i As Integer
        Dim j As Integer

        stataobj.DoCommand("set more off")
        stataobj.DoCommand("sysuse auto, clear")
     
        ' get number of observations and number of vars
        stataobj.DoCommand("describe")
        nobs = stataobj.StReturnNumeric("r(N)")
        nvar = stataobj.StReturnNumeric("r(k)")
        '
        ' get data of variables from 1 to 3. Since there are both 
        ' string and numeric variables, we need to use the variant
        ' array.
        ' 
        Dim varStart, varEnd As Integer
        Dim dataArr() As Object
        ' get data from variables from 1 to 3
        varStart = 1
        varEnd = 3
        dataArr = stataobj.VariablesData(1, 3, 1, nobs)
        ' display 
        For i = varStart To varEnd
            ' display var name
            varName = stataobj.StVariableName(i)
            Console.WriteLine(varName & vbCrLf)
            ' get varType
            varType = stataobj.VariableType(varName)
            If varType <= stata.StataVariableTypes.vstString And varType >= 1
            > Then
                ' string type
                Dim dataString As String()
                dataString = dataArr(i - varStart)
                ' display all data
                For j = 0 To nobs - 1
                    Console.WriteLine(j + 1 & ":" & dataString(j) & vbCrLf)
                Next
                Console.WriteLine(vbCrLf)
            End If

            If varType > stata.StataVariableTypes.vstString Then
                ' numeric type
                Dim dataDbl As Double()
                dataDbl = dataArr(i - varStart)
                ' display all data
                For j = 0 To nobs - 1
                    Console.WriteLine(j + 1 & ":" & dataDbl(j) & vbCrLf)
                Next
                Console.WriteLine(vbCrLf)
            End If
        Next
        stataobj = Nothing

3.14 Write Program using VBScript

According to Microsoft:

"The VBSCRIPT active scripting engine supplied by Microsoft only supports the indexing of SAFEARRAYs of VARIANTs. While VBSCRIPT is capable of accepting arrays of non-variant type for the purposes of boundary checking and passing it to other Automation objects, the engine does not allow manipulation of the array contents at this time."

Hence if a script attempts to reference the elements of an array returned by a method such as MatrixRowNames(), the script engine reports: "Object doesn't support this property or method stata.MatrixRowNames()".

In order to solve this problem, the following methods are provided for programming in VBScript. The method for VBScript returns an variant array instead of an array of specific data type comparing to the corresponding regular method.

Here is an example.

        [VBScript]

	' VBScript sample code
	dim stataobj
	set stataobj=CreateObject("stata.StataOLEApp")

	stataobj.DoCommand("sysuse auto")
	stataobj.DoCommand("regress mpg price")

	' get variable names as an array of variants
	dim varNames
	dim i
	varNames = stataobj.VariableNameAsVariants()
	For i=0 to 5
		msgbox "var " & i+1 & " " & varNames(i)
	Next
	
	' get variable labels as an array of variants
	dim varLabels 
	varLabels = stataobj.VariableLabelAsVariants()
	msgbox  "1st variable label =" & CStr(varLabels(0))

	' get the matrix column names as an array of variants
	dim colNames
	colNames = stataobj.MatrixRowNamesAsVariants("e(V)")
	For i=0 to 1 
		msgbox  "eV's " & i & "-th column name: " & CStr(colNames(i))
	Next

	' get matrix data as an array of variants.
	dim matentry
	matentry = stataobj.MatrixDataAsVariants("e(V)")
	For i=0 to 3
		msgbox  "eV's " & i & "-th = : " & matentry(i)
	Next
	
	' get value label -origin-'s values as an array of variants
	dim values  
	dim valcount
	valcount = stataobj.VariableValLabelCount("origin")
	values = stataobj.VariableValLabelValueAsVariants("origin")
	For i=0 to valcount-1 
		msgbox "Value label origin's " & i & "-th value=" & CInt(values(i))
	Next

	' get data as an array of variants
	dim make
	dim price
	make = stataobj.VariableDataAsVariants(1, 1, 5)
	price = stataobj.VariableDataAsVariants(2, 1, 5)
	For i=0 to 4
		msgbox  i+1 & "-th observation of variable make=" & make(i) 
		msgbox  i+1 & "-th observation of variable price=" & price(i)
	Next
	set stataobj=nothing

4. List of Stata Automation methods and properties