Update readstat

Fixes #235. Fixes #245. Closes #255.

NEWS.md

@@ -8,6 +8,7 @@
   * SAS: support Win baltic code page (#231)
   * SPSS: support uppercase time stamps (#230)
   * SPSS: fixes for 252-255 byte strings (#226)
+  * SPSS: fixes for 0 byte strings (#245)
 
 # haven 1.0.0
 

src/readstat/sas/readstat_sas7bdat_write.c

@@ -144,7 +144,18 @@ static readstat_error_t sas7bdat_emit_header(readstat_writer_t *writer, sas_head
     };
 
     memcpy(&header_start.magic, sas7bdat_magic_number, sizeof(header_start.magic));
-    strncpy(header_start.file_label, writer->file_label, sizeof(header_start.file_label));
+
+    memset(header_start.file_label, ' ', sizeof(header_start.file_label));
+
+    size_t file_label_len = strlen(writer->file_label);
+    if (file_label_len > sizeof(header_start.file_label))
+        file_label_len = sizeof(header_start.file_label);
+
+    if (file_label_len) {
+        memcpy(header_start.file_label, writer->file_label, file_label_len);
+    } else {
+        memcpy(header_start.file_label, "DATASET", sizeof("DATASET")-1);
+    }
 
     return sas_write_header(writer, hinfo, header_start);
 }

src/readstat/spss/readstat_sav_date.c

@@ -0,0 +1,81 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <math.h>
+
+static inline int is_leap(int year) {
+    return ((year % 4 == 0 && year % 100 != 0) || year % 400 ==0);
+}
+
+double readstat_sav_date_parse(const char *s, char **dest) {
+    // A SPSS date stored as the number of seconds since the start of the Gregorian calendar (midnight, Oct 14, 1582)
+    // Through the C interface in savReaderWriter I've verifed that leap seconds is ignored
+    int daysPerMonth[] =     {31,28,31,30,31,30,31,31,30,31,30,31};
+    int daysPerMonthLeap[] = {31,29,31,30,31,30,31,31,30,31,30,31};
+    int year, month, day;
+    if (strlen(s) == 0) {
+        *dest = (char*) s;
+        return 0;
+    }
+    int ret = sscanf(s, "%d-%d-%d", &year, &month, &day);
+    month--;
+    if (month < 0 || month > 11 || ret!=3) {
+        *dest = (char*)s;
+        return 0;
+    }
+    int maxdays = (is_leap(year) ? daysPerMonthLeap : daysPerMonth)[month]; 
+    if (day < 1 || day > maxdays) {
+        *dest =(char*)s;
+        return 0;
+    } else {
+        int days = 0;
+
+        for (int i=1582; i<year; i++) {
+            days += is_leap(i) ? 366 : 365;
+        }
+
+        for (int m=0; m<month; m++) {
+            days += is_leap(year) ? daysPerMonthLeap[m] : daysPerMonth[m];
+        }
+
+        days += day-1;
+        char buf[1024];
+        *dest = (char*)s + snprintf(buf, sizeof(buf), "%d-%d-%d", year, month+1, day); 
+        return (days * 86400.0) - 24710400; // 24710400 is the number of seconds in 1582 before Oct 14
+    }
+}
+
+char* readstat_sav_date_string(double seconds, char* dest, int size) {
+    int yr = 1582;
+    int month = 0;
+    int day = 1;
+    int daysPerMonth[] =     {31,28,31,30,31,30,31,31,30,31,30,31};
+    int daysPerMonthLeap[] = {31,29,31,30,31,30,31,31,30,31,30,31};
+    double secs = seconds;
+    secs += 24710400.0;
+    double days = secs / 86400.0;
+    double err = ceil(days) - days;
+    if (err != 0.0) {
+        fprintf(stderr, "%s:%d time not supported. seconds was %lf, err was %lf\n", __FILE__, __LINE__, seconds, err);
+        return NULL;
+    }
+
+    while (days > 0) {
+        int days_in_year = is_leap(yr) ? 366 : 365;
+        if (days >= days_in_year) {
+            yr+=1;
+            days-=days_in_year;
+            continue;
+        }
+        int days_in_month = is_leap(yr) ? daysPerMonthLeap[month] : daysPerMonth[month];
+        if (days >= days_in_month) {
+            month+=1;
+            days-=days_in_month;
+            continue;
+        }
+        day+= days;
+        days = 0;
+    }
+    snprintf(dest, size, "%04d-%02d-%02d", yr, month+1, day);
+    return dest;
+}

src/readstat/spss/readstat_sav_date.h

@@ -0,0 +1,7 @@
+#ifndef __READSTAT_SAV_DATE_H
+#define __READSTAT_SAV_DATE_H
+
+double readstat_sav_date_parse(const char *s, char **dest);
+char* readstat_sav_date_string(double seconds, char* dest, int size);
+
+#endif

src/readstat/spss/readstat_sav_write.c

@@ -907,7 +907,9 @@ static size_t sav_variable_width(readstat_type_t type, size_t user_width) {
             size_t last_segment_width = ((user_width - (n_segments - 1) * 252) + 7)/8*8;
             return (n_segments-1)*256 + last_segment_width;
         }
-
+        if (user_width == 0) {
+            return 8;
+        }
         return (user_width + 7) / 8 * 8;
     }
     return 8;

src/readstat/stata/readstat_dta.h

@@ -11,11 +11,19 @@ typedef struct dta_header_s {
     int32_t          nobs;
 } dta_header_t;
 
-typedef struct dta_strl_header_s {
-    unsigned char   vo_bytes[8];
+typedef struct dta_117_strl_header_s {
+    uint32_t        v;
+    uint32_t        o;
     unsigned char   type;
     int32_t         len;
-} dta_strl_header_t;
+} dta_117_strl_header_t;
+
+typedef struct dta_118_strl_header_s {
+    uint32_t        v;
+    uint64_t        o;
+    unsigned char   type;
+    int32_t         len;
+} dta_118_strl_header_t;
 
 #pragma pack(pop)
 

src/readstat/stata/readstat_dta_days.c

@@ -0,0 +1,96 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+static inline int is_leap(int year) {
+    return ((year % 4 == 0 && year % 100 != 0) || year % 400 ==0);
+}
+
+int readstat_dta_num_days(const char *s, char **dest) {
+    int daysPerMonth[] =     {31,28,31,30,31,30,31,31,30,31,30,31};
+    int daysPerMonthLeap[] = {31,29,31,30,31,30,31,31,30,31,30,31};
+    int year, month, day;
+    if (strlen(s) == 0) {
+        *dest = (char*) s;
+        return 0;
+    }
+    int ret = sscanf(s, "%d-%d-%d", &year, &month, &day);
+    month--;
+    if (month < 0 || month > 11 || ret!=3) {
+        *dest = (char*)s;
+        return 0;
+    }
+    int maxdays = (is_leap(year) ? daysPerMonthLeap : daysPerMonth)[month]; 
+    if (day < 1 || day > maxdays) {
+        *dest = (char*)s;
+        return 0;
+    } else {
+        int days = 0;
+
+        for (int i=year; i<1960; i++) {
+            days -= is_leap(i) ? 366 : 365;
+        }
+
+        for (int i=1960; i<year; i++) {
+            days += is_leap(i) ? 366 : 365;
+        }
+
+        for (int m=0; m<month; m++) {
+            days += is_leap(year) ? daysPerMonthLeap[m] : daysPerMonth[m];
+        }
+
+        days += day-1;
+        char buf[1024];
+        *dest = (char*)s + snprintf(buf, sizeof(buf), "%d-%d-%d", year, month+1, day); 
+        return days;
+    }
+}
+
+char* readstat_dta_days_string(int days, char* dest, int size) {
+    // TODO: Candidate for clean up
+    int yr = 1960;
+    int month = 0;
+    int day = 1;
+    int daysPerMonth[] =     {31,28,31,30,31,30,31,31,30,31,30,31};
+    int daysPerMonthLeap[] = {31,29,31,30,31,30,31,31,30,31,30,31};
+    if (days < 0) {
+        yr = 1959;
+        month = 11;
+        days = - days;
+        while (days > 0) {
+            int days_in_year = is_leap(yr) ? 366 : 365;
+            if (days > days_in_year) {
+                yr-=1;
+                days-=days_in_year;
+                continue;
+            }
+            int days_in_month = is_leap(yr) ? daysPerMonthLeap[month] : daysPerMonth[month];
+            if (days > days_in_month) {
+                month-=1;
+                days-=days_in_month;
+                continue;
+            }
+            day = days_in_month-days + 1;
+            days = 0;
+        }
+    } else {
+        while (days > 0) {
+            int days_in_year = is_leap(yr) ? 366 : 365;
+            if (days >= days_in_year) {
+                yr+=1;
+                days-=days_in_year;
+                continue;
+            }
+            int days_in_month = is_leap(yr) ? daysPerMonthLeap[month] : daysPerMonth[month];
+            if (days >= days_in_month) {
+                month+=1;
+                days-=days_in_month;
+                continue;
+            }
+            day+= days;
+            days = 0;
+        }
+    }
+    snprintf(dest, size, "%04d-%02d-%02d", yr, month+1, day);
+    return dest;
+}

src/readstat/stata/readstat_dta_days.h

@@ -0,0 +1,7 @@
+#ifndef __READSTAT_DTA_DAYS_H
+#define __READSTAT_DTA_DAYS_H
+
+int readstat_dta_num_days(const char *s, char** dest);
+char* readstat_dta_days_string(int days, char* dest, int size);
+
+#endif