根据stata官网的信息，stata13已经更新至2015年06月03日：
http://www.stata.com/support/updates/stata13.html



多谢版主提醒：【非正版用户】请务必使用【离线更新】方式：
1、在上述网址链接中下载相应版本的更新压缩包，并解压缩到你选定的某个硬盘地址
2、在stata13的命令窗口中输入命令：
db update
3、然后按照弹窗提示将更新源指定为上述解压缩文件夹所在的硬盘地址


-------- update 03jun2015 -------------------------------------------------------------------------------------------------

     1.  ereturn display, ml display, and _coef_table, when reporting estimation results from a multiple-equation model
         that included an interaction term in more than one equation, would sometimes fail to recognize some base levels
         of interaction terms.  In this case, base levels were labeled "(omitted)" instead of "(base)".  The estimated
         coefficients and related statistics were correct; the error affected only the labels.

         For example, suppose we have factors A and B, each with levels 1 and 2.  Then the coefficient table from

                . sureg (y1 = A##B) (y2 = A#B), allbase

         would mislabel elements _b[y1:1.A#2.B] and _b[y1:2.A#1.B] as "(omitted)" instead of as "(base)".  This has been
         fixed.

     2.  estat mfx, after asclogit or asmprobit with a value label assigned to varname specified in alternatives(),
         exited with an error if any strings in the value label were longer than 26 characters.  This has been fixed.

     3.  Any estimation command that allowed factor variables would slow down when matsize was set larger than 1,000 even
         for models with relatively few parameters.  This has been fixed.

     4.  export excel, when exporting variables with any calendar date format other than %td, incorrectly exported the
         variables as a daily date instead of respecting the weekly, monthly, quarterly, half-yearly, or yearly format.
         This has been fixed.

     5.  For variable names of the form d#, D#, e#, and E#, Stata's expression parser misinterpreted some factor-variable
         specifications such as 1.d1 and 1.e2 as numeric literals instead of the intended indicator variables.  This has
         been fixed.

     6.  Functions normalden(x,m,s) and lnnormalden(x,m,s), when s was specified as a string value rather than as a
         numeric value representing the standard deviation, returned the specified string instead of exiting with an
         error message.  This has been fixed.

     7.  gnbreg, nbreg, and tnbreg now report the pseudo-R-squared when option vce(robust), option vce(cluster clustvar),
         or pweights are specified.

     8.  graph, when using a .gph file that was created by a newer version of Stata that required a newer format .gph
         file and when the graph file name was quoted on the command line, simply declared the file to be not found.
         graph now issues the message that the format of the file is too new for the running version of Stata.

     9.  graph bar, when no variable or statistic was specified and when options over() and by() were combined (for
         example, graph bar, over(rep78) by(foreign)), incorrectly exited with an error.  This has been fixed.

    10.  graph pie, when the number of slices exceeded the defined number of pstyles in a scheme, made all excess slices
         white and issued uninformative notes.  graph pie now behaves like graph bar, graph twoway, and other graph
         commands by recycling the defined pstyles when the number of slices exceeds the number of defined pstyles.
         graph pie now also supports the pcycle() option to reduce the number of pstyles used before recycling.

    11.  import haver has the following fixes:

         a.  import haver, when aggregating a daily series to weekly, defined weeks as beginning on Mondays.  Friday is
             now treated as the first day of the week.

         b.  import haver, when a series contained no data, loaded the series as a variable containing all missing values
             instead of dropping the series from the query.  This has been fixed.

         c.  import haver, describe with option saving() always displayed the series meta-information in the Results
             window.  The meta-information should only be displayed when suboption verbose is specified in saving().
             This has been fixed.

    12.  irf graph with option ci#opts() did not change the rendition of the confidence interval for the #th statistic.
         This has been fixed.

    13.  ivprobit for models with more than one endogenous variable specified is now more likely to converge.

    14.  margins after me commands with one or more factor variables that have their base level fvset incorrectly exited
         with an error.  This has been fixed.

    15.  mi impute intreg exited with a Mata error "st_store():  3203 colvector required" when incomplete data contained
         only one censored observation.  This has been fixed.

    16.  mi impute pmm, mi impute monotone, and mi impute chained now display a note whenever predictive mean matching
         imputation uses the default one nearest neighbor.  This default is arbitrary.  The optimal number of nearest
         neighbors varies from one application to another.  Recent simulation studies demonstrated that using one nearest
         neighbor performed poorly in many of the considered scenarios.  We suggest that you choose the number of
         neighbors appropriate for your data and specify it in option knn() when imputing variables using predictive mean
         matching.

    17.  minbound, when the minimum value of the function occurred at the specified left boundary, returned the function
         value evaluated close to, but not exactly at, the boundary.  This has been fixed.

    18.  ml and Mata function moptimize(), with lf1 specified as the evaluator type, were passing the value 2 instead of
         1 for todo while computing numerical second derivatives.  This has been fixed.

    19.  mlexp ignored reporting options associated with factor variables.  This has been fixed.

    20.  mvreg, when replaying results from a model fit using manova, did not display value labels for factor variables.
         This has been fixed.

    21.  pause would become stuck in an endless loop if you tried to start an interactive Mata session while Stata was
         already in a paused state.  pause now returns an error message in this case.

    22.  predict with option pr() after etpoisson did not correctly compute the predicted probabilities.  predict failed
         to integrate over the unobserved component to obtain the final results.  This has been fixed.

    23.  predict with option lnsigma after teffects ra, when the specified prediction created one new variable instead of
         a group of variables, incorrectly exited with an error.  This has been fixed.

    24.  set coeftabresults off is a setting used to make regress faster when run in tight loops, in simulations, and
         when bootstrapping.  It is now faster still.

         Other estimation commands are also affected, but you would rarely notice a difference.

    25.  regress, when a time-series-operated variable was omitted because of collinearity, reported that a variable
         named _delete was omitted instead of reporting the name of the omitted variable.  This has been fixed.

    26.  sem has the following fixes:

         a.  sem, when fitting a model with no observed exogenous variables, with more than one latent variable, and with
             at least two coefficients constrained to 1 on paths from a latent variable to its observed measurements,
             incorrectly exited with an error message.  This has been fixed.

         b.  sem, when option covariance() specified more than one covariance term and when starting values were
             specified for a subset of those covariance terms via the init() option, sometimes incorrectly exited with an
             error message.  Whether an error was reported depended on the order of the terms within the covariance()
             option.  For example, the option covariance(x1*x2 (x1*x3, init(.2))) would trigger this error.  This has
             been fixed.

         c.  In the SEM Builder, canvas length and canvas width were not properly restricted to be between 1 inch and 20
             inches.  This has been fixed.

    27.  Stata matrix row and column names parsing logic, when presented with interactions where the order of factor
         variables was not the same in each term, did not properly keep track of the levels.  For example,

             . matrix colna b = 0.foreign#1.rep78 2.rep78#1.foreign

         would result in colnames

             . matrix colna b = 0.foreign#1.rep78 2.foreign#1.rep78

         instead of the intended

             . matrix colna b = 0.foreign#1.rep78 1.foreign#2.rep78

         This has been fixed.

    28.  stcurve, when plotting curves for a model that included an interaction between a factor variable and a
         continuous variable and when option at() set the continuous variable to a value that was not observed in the
         data, incorrectly reported error message "# is not a valid level for factor variable varname".  This has been
         fixed.

    29.  Time-series operators have the following fixes:

         a.  Seasonal difference time-series operator S# did not appropriately translate the zero-period difference (S0).
             Depending on the command with which it was specified, S0 was translated to S1 or resulted in a variable
             containing only 0 values.  In all cases, typing S0.var is now equivalent to typing var; this translation is
             the same as that used for the other time-series operators.

         b.  Time-series operators, when used with a dataset with more than 1 billion observations and when the time
             series contained gaps, could cause Stata to crash or to return incorrect results.  Whether this occurred or
             not depended on certain relationships between the location of the gap, the number of observations, and the
             operator used.  This has been fixed.

    30.  twoway scatter, twoway line, twoway connected, twoway area, twoway bar, twoway spike, twoway dropline, twoway
         dot, twoway rarea, twoway rbar, twoway rspike, twoway rcap, twoway rcapsym, twoway rscatter, twoway rline, and
         twoway rconnected now support directly graphing the columns of a Mata matrix.  For the sake of speed and space,
         the matrix data are passed directly to the graph, not brought into the Stata dataset.

         The basic syntax is

             twoway scatter matamatrix(matname) [namelist]

         This feature is not officially documented, but the syntax is documented in help twoway mata.

    31.  xtgls issued a note when the number of panels was greater than the number of periods.  The note incorrectly
         stated the opposite, that the number of periods was greater than the number of panels.  This has been fixed.

    32.  xtivreg with option vce(bootstrap) or option vce(jackknife) exited with an error when an independent variable
         was included after the endogenous variable and instruments were specified in parentheses.  For example, the
         placement of variable not_smsa in

             . xtivreg ln_wage c.age##c.age (tenure = union south) not_smsa, vce(bootstrap)

         would trigger this error.  Variables can now be specified after the expression in parentheses.

    33.  Executing an operation using the https:// protocol when an internal content buffer was less than 100 bytes would
         cause Stata to exit with return code r(1).  This typically happened only when the results to be returned by the
         operation were very short.  This has been fixed.

    34.  (64-bit platforms) Mata functions fseek() and ftell(), when used to search for a position larger than 2GB in a
         file that was also larger than 2GB, returned a negative error code.  This has been fixed.

    35.  (Mac) import excel, when importing long variable names, incorrectly truncated the variable names to 33
         characters when it should have truncated them to 32 characters.  This has been fixed.

    36.  (Mac) translate with translator smcl2pdf and when the destination file was specified either without an absolute
         path or without a path beginning ~/ resulted in an empty destination file.  This has been fixed.

    37.  (Mac) xshell did not launch xterm when used on Mac OS X 10.8 or newer because XQuartz installs the xterm
         executable in a location other than where Stata is expecting.  This has been fixed.

    38.  (Mac) The Do-file Editor has the following fixes:

         a.  Line numbers were not correctly resized when the Do-file Editor was zoomed in or zoomed out.  This has been
             fixed.

         b.  Using the Find dialog and selecting "Replace all" might not replace all the text specified in "Find what:"
             if 1) the text to be replaced occurs more than once; 2) the length of the text specified in "Replace what:"
             is longer than the text that is being replaced; and 3) the text that is being replaced comes very close to
             the end of the document.

    39.  (Mac) Stata graphics no longer attempts to render a 0 line width path but will instead ignore the drawing
         command.  In addition, Stata no longer sets the minimum line width to 1 pixel when rendering paths on the
         screen.  Users with Mac OS X may notice that there is no longer any variation in how 0 width lines are rendered
         across output devices such as monitors and printers.

    40.  (Unix GUI) If a user changed themes on the computer, the Preferences dialog box continued to use a black font
         even if the theme requested another color.  This has been fixed.

    41.  (Unix GUI) The SEM Builder's Estimation menu, when used to request exponentiated coefficients after fitting a
         generalized SEM, opened an incorrect dialog box.  This has been fixed.

    42.  (Linux/Unix) In the Stata class system, if Stata was launched in batch mode, destructors were not invoked when
         Stata exited.  This affected only users who had created classes that used destructors to manage computing
         resources not internal to Stata.  This has been fixed.