missing

missing

Functions

#define na()
#define isfinite()
#define xna()
#define model_missing()
int model_has_missing_obs ()
int first_missing_index ()
int series_adjust_sample ()
int list_adjust_sample ()
int set_miss ()
double missing_obs_fraction ()
int any_missing_user_values ()

Types and Values

#define NEW_NA
#define NADBL

Object Hierarchy


Description

Functions

na()

#define na(x) ((x) == NADBL)


isfinite()

# define isfinite(x) (!isnan(x) && !isinf(x))


xna()

# define xna(x) ((x) == NADBL || isnan(x) || isinf(x))


model_missing()

#define model_missing(m,t) ((m)->missmask != NULL && (m)->missmask[t] == '1')


model_has_missing_obs ()

int
model_has_missing_obs (const MODEL *pmod);

Parameters

pmod

pointer to model.

 

Returns

1 if there are missing observations in the model's sample range, otherwise 0.


first_missing_index ()

int
first_missing_index (const double *x,
                     int t1,
                     int t2);

Parameters

x

array to be checked for missing values.

 

t1

start of range to check.

 

t2

end of range to check.

 

Returns

the index of the first missing observation in x over the sample range t1 to t2 , or -1 if there is no such observation.


series_adjust_sample ()

int
series_adjust_sample (const double *x,
                      int *t1,
                      int *t2);

Adjusts t1 and t2 so as to drop any leading or trailing missing observations.

Parameters

x

series to be checked for missing values.

 

t1

on entry, initial start of sample range; on exit, start of sample range adjusted for missing values.

 

t2

on entry, initial end of sample range; on exit, end of sample range adjusted for missing values.

 

Returns

E_MISSDATA if interior missing values were found within the (possibly adjusted) sample range, otherwise 0.


list_adjust_sample ()

int
list_adjust_sample (const int *list,
                    int *t1,
                    int *t2,
                    const DATASET *dset,
                    int *nmiss);

Drops leading or trailing observations from the sample range initially given by the values in t1 and t2 if missing values are found for any of the variables given in list .

If nmiss is non-NULL it receives the number of missing values inside the (possibly reduced) sample range, otherwise it is considered an error if there are any such missing values.

Parameters

list

list of variables to be tested for missing values.

 

t1

on entry, initial start of sample range; on exit, start of sample range adjusted for missing values.

 

t2

on entry, initial end of sample range; on exit, end of sample range adjusted for missing values.

 

dset

dataset struct.

 

nmiss

location to receive number of missing values within (possibly adjusted) sample range.

 

Returns

0 on success or E_MISSDATA or error.


set_miss ()

int
set_miss (const int *list,
          const char *param,
          DATASET *dset,
          PRN *prn);

Set to "missing" each observation of each series in list that has the value represented by param .

Parameters

list

list of variables to process, or an empty list or NULL to process all variables.

 

param

string representation of the numerical value to treat as missing.

 

dset

dataset struct.

 

prn

pointer to printing struct.

 

Returns

1 if at least one observation was set as missing, otherwise 0.


missing_obs_fraction ()

double
missing_obs_fraction (const DATASET *dset);

Parameters

dset

dataset struct.

 

Returns

the fraction of the observations in Z for which all variables have missing values (empty rows).


any_missing_user_values ()

int
any_missing_user_values (const DATASET *dset);

Parameters

dset

dataset struct.

 

Returns

1 if there are missing values for any non-hidden variables within the current sample range, otherwise 0.

Types and Values

NEW_NA

#define NEW_NA 0


NADBL

#define NADBL DBL_MAX