whatsnew7to8.hlp

是一个经济学管理应用软件 很难找的 但是经济学学生又必须用到

    use whichever appeals to you.

{p 10 10 2}
    New function {cmd:mi()} is a synonym for existing function
    {cmd:missing()}; it returns 1 (true) if missing and false otherwise.
    See help {help progfun}.

{p 6 10 2}
2.  By the same token, do not type

{p 14 18 2}
	{cmd:. list} ... {cmd:if} {it:var} {cmd:== .}

{p 10 10 2}
    To list observations with missing values of {\it var}, type

{p 14 18 2}
	{cmd:. list} ... {cmd:if} {it:var} {cmd:>= .}

{p 10 10 2}
    or type

{p 14 18 2}
	{cmd:. list} ... {cmd:if mi(}{it:var}{cmd:)}

{p 6 10 2}
3.  Matrices can now contain missing values, both the standard one ({cmd:.})
    and the extended ones ({cmd:.a}, {cmd:.b}, ..., {cmd:.z}).

{p 6 10 2}
4.  The following new density functions are provided:

{p 10 14 2}
    a.  {cmd:tden(}{it:n}{cmd:,}{it:t}{cmd:)}, the density of Student's t
	distribution.

{p 10 14 2}
    b.  {cmd:Fden(}{it:n_1}{cmd:,}{it:n_2}{cmd:,}{it:F}{cmd:)}, the density of
	the F distribution.

{p 10 14 2}
    c.  {cmd:nFden(}{it:n_1}{cmd:,}{it:n_2}{cmd:,}{it:lambda}{cmd:,}{it:F}{cmd:)},
	the noncentral F density.

{p 10 14 2}
    d.  {cmd:betaden(}{it:a}{cmd:,}{it:b}{cmd:,}{it:x}{cmd:)}, the 2-parameter
	Beta density.

{p 10 14 2}
    e.  {cmd:nbetaden(}{it:a}{cmd:,}{it:b}{cmd:,}{it:g}{cmd:,}{it:x}{cmd:)},
	the noncentral Beta density.

{p 10 14 2}
    f.  {cmd:gammaden(}{it:a}{cmd:,}{it:b}{cmd:,}{it:g}{cmd:,}{it:x}{cmd:)},
	the 3-parameter Gamma density.

{p 10 10 2}
    See help {help probfun}.

{p 6 10 2}
5.  The following new cumulative density functions are provided:

{p 10 14 2}
    a.  {cmd:nFtail(}{it:n_1}{cmd:,}{it:n_2}{cmd:,}{it:lambda}{cmd:,}{it:f}{cmd:)},
	the upper-tail of the noncentral F.

{p 10 14 2}
    b.  {cmd:nibeta(}{it:a}{cmd:,}{it:b}{cmd:,}{it:lambda}{cmd:,}{it:x}{cmd:)},
	the cumulative noncentral ibeta probability.

{p 10 10 2}
    See help {help probfun}.

{p 6 10 2}
6.  The following new inverse cumulative density functions are provided:

{p 10 14 2}
    a.  {cmd:invnFtail(}{it:n_1}{cmd:,}{it:n_2}{cmd:,}{it:lambda}{cmd:,}{it:p}{cmd:)},
	the noncentral F corresponding to upper-tail {it:p}.

{p 10 14 2}
    b.  {cmd:invibeta(}{it:a}{cmd:,}{it:b}{cmd:,}{it:p}{cmd:)}, the incomplete
	beta value corresponding to {it:p}.

{p 10 14 2}
    c.  {cmd:invnibeta(}{it:a}{cmd:,}{it:b}{cmd:,}{it:lambda}{cmd:,}{it:p}{cmd:)},
	the noncentral beta value corresponding to {it:p}.

{p 10 10 2}
    In addition, existing function
    {cmd:invbinomial(}{it:n}{cmd:,}{it:k}{cmd:,}{it:p}{cmd:)} has improved
    accuracy.  See help {help probfun}.

{p 6 10 2}
7.  A suite of new functions provides partial derivatives of the cumulative
    gamma distribution.  The following new functions are provided:

{p 10 14 2}
    a.  {cmd:dgammapda(}{it:a}{cmd:,}{it:x}{cmd:)}, partial derivative of
	{cmd:gammap(}{it:a}{cmd:,}{it:x}{cmd:)} with respect to {it:a}.

{p 10 14 2}
    b.  {cmd:dgammapdx(}{it:a}{cmd:,}{it:x}{cmd:)}, partial derivative of
	{cmd:gammap(}{it:a}{cmd:,}{it:x}{cmd:)} with respect to {it:x}.

{p 10 14 2}
    c.  {cmd:dgammapdada(}{it:a}{cmd:,}{it:x}{cmd:)}, 2nd partial derivative
	of {cmd:gammap(}{it:a}{cmd:,}{it:x}{cmd:)} with respect to {it:a}.

{p 10 14 2}
    d.  {cmd:dgammapdxdx(}{it:a}{cmd:,}{it:x}{cmd:)}, 2nd partial derivative
	of {cmd:gammap(}{it:a}{cmd:,}{it:x}{cmd:)} with respect to {it:x}.

{p 10 14 2}
    e.  {cmd:dgammapdadx(}{it:a}{cmd:,}{it:x}{cmd:)}, 2nd partial derivative
	of {cmd:gammap(}{it:a}{cmd:,}{it:x}{cmd:)} with respect to {it:a} and
	{it:x}.

{p 10 10 2}
    See help {help probfun}.

{p 6 10 2}
8.  All density and distribution functions have been extended to return
    nonmissing values over the entire real line; see help {help probfun}.

{p 6 10 2}
9.  The following new string functions are provided:

{p 10 14 2}
    a.  {cmd:word(}{it:s}{cmd:,}{it:n}{cmd:)} returns the {it:n}th word in
	{it:s}.

{p 10 14 2}
    b.  {cmd:wordcount(}{it:s}{cmd:)} returns the number of words in {it:s}.

{p 10 14 2}
    c.  {cmd:char(}{it:n}{cmd:)} returns the character corresponding to ASCII
	code {it:n}.

{p 10 14 2}
    d.  {cmd:plural(}{it:n}{cmd:,}{it:s_1}{cmd:)} returns the plural of
	{it:s_1} if {it:n} does not equal 1 or -1, and otherwise returns
	{it:s_1}.

{p 10 14 2}
    e.  {cmd:plural(}{it:n}{cmd:,}{it:s_1}{cmd:,}{it:s_2}{cmd:)} returns the
	plural of {it:s_1} if {it:n} does not equal 1 or -1, forming the
	plural by adding or removing suffix {cmd:s_2}.

{p 10 14 2}
    f.  {cmd:proper(}{it:s}{cmd:)} capitalizes the first letter of a string
	and any other letters immediately following characters that are not
	letters; remaining letters are converted to lowercase.

{p 10 10 2}
See help {help strfun}.

{p 5 10 2}
10.  The following new mathematical functions are provided:

{p 10 14 2}
     a.  {cmd:logit(}{it:x}{cmd:)}, the log of the odds ratio.

{p 10 14 2}
     b.  {cmd:invlogit(}{it:x}{cmd:)}, the inverse logit.

{p 10 14 2}
     c.  {cmd:cloglog(}{it:x}{cmd:)}, the complementary log-log.

{p 10 14 2}
     d.  {cmd:invcloglog(}{it:x}{cmd:)}, the inverse of the complementary
	 log-log.

{p 10 14 2}
     e.  {cmd:tanh(}{it:x}{cmd:)}, the hyperbolic tangent.

{p 10 14 2}
     f.  {cmd:atanh(}{it:x}{cmd:)}, the arc-hyperbolic tangent of {it:x}.

{p 10 14 2}
     g.  {cmd:floor(}{it:x}{cmd:)}, the integer {it:n} such that
	 {it:n} <= {it:x} < {it:n}+1.

{p 10 14 2}
     h.  {cmd:ceil(}{it:x}{cmd:)}, the integer {it:n} such that
	 {it:n} < {it:x} <= {it:n}+1.

{p 10 10 2}
    In addition, the following existing mathematical functions have been
    modified:

{p 10 14 2}
     i.  {cmd:round(}{it:x}{cmd:,}{it:y}{cmd:)} now allows the second argument
	 be optional and defaults it to 1, so {cmd:round(}{it:x}{cmd:)}
	 returns {it:x} rounded to the closest integer.

{p 10 14 2}
     j.  {cmd:lngamma(}{it:x}{cmd:)} and
	 {cmd:gammap(}{it:a}{cmd:,}{it:x}{cmd:)} now have improved accuracy.

{p 10 10 2}
See help {help mathfun}.

{p 5 10 2}
11.  Existing function {cmd:uniform()} will now allow you to capture and reset
     its seed.  The seed value, in encrypted form, is now shown by
     {cmd:query}.  You can store its value by typing

{p 14 18 2}
	{cmd:local seed = c(seed)}

{p 10 10 2}
     Later, you can reset it by typing

{p 14 18 2}
	{cmd:. set seed `seed'}

{p 10 10 2}
    See help {help seed} and help {help random}.

{p 5 10 2}
12.  The following new matrix functions are provided:

{p 10 14 2}
     a.  {cmd:issym(}{it:M}{cmd:)} returns 1 if matrix {it:M} is symmetric and
	 returns 0 otherwise; {cmd:issym()} may be used in any context.

{p 10 14 2}
     b.  {cmd:matmissing(}{it:M}{cmd:)} returns 1 if any elements of {it:M}
	 are missing and returns 0 otherwise; {cmd:matmissing()} may be used
	 in any context.

{p 10 14 2}
     c.  {cmd:vec(}{it:M}{cmd:)} returns the column vector formed by listing
	 the elements of {it:M}, starting with the first column and proceeding
	 column by column.

{p 10 14 2}
     d.  {cmd:hadamard(}{it:M}{cmd:,}{it:N}{cmd:)} returns a matrix whose
	 {it:i}, {it:j} element is
	 {it:M}[{it:i},{it:j}] * {it:N}[{it:i},{it:j}].

{p 10 14 2}
     e.  {cmd:matuniform(}{it:r}{cmd:,}{it:c}{cmd:)} returns the {it:r} by
	 {it:c} matrix containing uniformly distributed pseudo-random numbers
	 on the interval [0,1).

{p 10 10 2}
     See help {help matfcns}.

{p 10 10 2}
     In addition, the new command {cmd:matrix} {cmd:eigenvalues} returns the
     complex eigenvalues of an {it:n} by {it:n} nonsymmetric matrix; see help
     {help mateig}.

{p 5 10 2}
13.  The following new programming functions have been added:

{p 10 14 2}
     a.  {cmd:clip(}{it:x}{cmd:,}{it:a}{cmd:,}{it:b}{cmd:)} returns {it:x} if
	 {it:a} <= {it:x} <= {it:b}, {it:a} if {it:x} <= {it:a}, {it:b} if
	 {it:x} >= {it:b}, and {it:missing} if $x$ is missing.

{p 10 14 2}
     b.  {cmd:chop(}{it:x}{cmd:,}{it:epsilon}{cmd:)} returns
	 {cmd:round(}{it:x}{cmd:)} if
	 |{it:x} - {cmd:round(}{it:x}{cmd:)}| < {it:epsilon}, otherwise
	 returns {it:x}.

{p 10 14 2}
     c.  {cmd:irecode(}{it:z}{cmd:,}{it:x_1}{cmd:,}{it:x_2}{cmd:,} ... {cmd:,}{it:x_n}{cmd:)}
	 returns the index of the range in which {it:z} falls.

{p 10 14 2}
     d.  {cmd:maxbyte()}, {cmd:maxint()}, {cmd:maxlong()},
	 {cmd:maxfloat()}, and {cmd:maxdouble()} return the maximum value
	 allowed by the storage type.

{p 10 14 2}
     e.  {cmd:minbyte()}, {cmd:minint()}, {cmd:minlong()},
	 {cmd:minfloat()}, and {cmd:mindouble()} return the minimum value
	 allowed by the storage type.

{p 10 14 2}
     f.  {cmd:epsfloat()} and {cmd:epsdouble()} return the precision
	 associated with the storage type.

{p 10 14 2}
     g.  {cmd:byteorder()} returns 1 if the computer stores numbers in
	 most-significant-byte-first format and 0 if in
	 least-significant-byte-first format.

{p 10 10 2}
     The following programming functions have been modified or extended:

{p 10 14 2}
     h.  {cmd:missing(}{it:x}{cmd:)} now optionally allows multiple arguments
	 so that it becomes
	 {cmd:missing(}{it:x_1}{cmd:,}{it:x_2}{cmd:,} ... {cmd:,}{it:x_n}{cmd:)}.
	 The extended function returns 1 (true) if any of the {it:x_i} are
	 missing and returns 0 (false) otherwise.

{p 10 14 2}
     i.  {cmd:cond(}{it:x}{cmd:,}{it:a}{cmd:,}{it:b}{cmd:)} now optionally
	 allows a fourth argument so that it becomes
	 {cmd:cond(}{it:x}{cmd:,}{it:a}{cmd:,}{it:b}{cmd:,}{it:c}{cmd:)}.
	 {it:c} is returned if {it:x} evaluates to missing.

{p 10 10 2}
See help {help progfun}.


{title:What's new in display formats}

{p 6 10 2}
1.  The {cmd:%g} format has been modified:  {cmd:%}{it:#}{cmd:.0g} still means
    the same as previously, but {cmd:%}{it:#}{cmd:.}{it:#}{cmd:g} has a new
    meaning.  For instance, {cmd:%9.5g} means to show approximately 5
    significant digits.  We say approximately because, given the number
    123,456, {cmd:%9.5g} will show {cmd:123456} rather than {cmd:1.2346e+05},
    as would strictly be required if only five digits are to be shown.  Other
    than that, it does what you would expect, and we think, in all cases, does
    what you want.

{p 6 10 2}
2.  {cmd:%}[{cmd:-}]{cmd:0}{it:#}{cmd:.}{it:#}{cmd:f} formats, note the
    leading {cmd:0}, now specify that leading zeros are to be included in the
    result.  1.2 in {cmd:%09.2f} format is {cmd:000001.20}.

{p 6 10 2}
3.  Stata has a new {cmd:%21x} hexadecimal format that will mainly be of
    interest to numerical analysts.  In {cmd:%21x}, 123,456 looks like
    {cmd:+1.e240000000000X+010}, which you read as the hexadecimal number
    1.e24 multiplied by 2^10.  The period in 1.e24 is the base-16 point.
    The beauty of this format is that it reveals numbers exactly as the
    binary computer thinks of it.  For instance, the new format shows how
    difficult numbers like 0.1 are for binary computers:
    {cmd:+1.999999999999aX-004}.

{p 10 10 2}
    You can use this hexadecimal way of writing numbers in expressions; Stata
    will understand, for instance,

{p 14 18 2}
	{cmd:. generate xover4 = x / 1.0x+2}

{p 10 10 2}
    but it is unlikely you would want to do that.  The notation will even by
    understood by {help input}, {help infix}, and {help infile}.  There is no
    {cmd:%21x} input format, but wherever a number appears, Stata will
    understand {it:#}{cmd:.}{it:##}..