index 4d7fda31c11d590a865fc6aff0665292249f0014..c179fea84b142f836a24eb4791eb12194344f12e 100644 (file)
--- a/src/include/core/data.h
+++ b/src/include/core/data.h
#include "core/time.h"
#include <inttypes.h>
+#include <stdbool.h>
#include <stddef.h>
+#include <sys/types.h>
+#include <regex.h>
+
#ifdef __cplusplus
extern "C" {
#endif
enum {
- SDB_TYPE_INTEGER = 1,
+ SDB_TYPE_NULL = 0,
+ SDB_TYPE_INTEGER,
SDB_TYPE_DECIMAL,
SDB_TYPE_STRING,
SDB_TYPE_DATETIME,
SDB_TYPE_BINARY,
+ SDB_TYPE_REGEX, /* extended, case-insensitive POSIX regex */
+
+ /* flags: */
+ SDB_TYPE_ARRAY = 1 << 8,
};
#define SDB_TYPE_TO_STRING(t) \
- (((t) == SDB_TYPE_INTEGER) \
- ? "INTEGER" \
- : ((t) == SDB_TYPE_DECIMAL) \
- ? "DECIMAL" \
- : ((t) == SDB_TYPE_STRING) \
- ? "STRING" \
- : ((t) == SDB_TYPE_DATETIME) \
- ? "DATETIME" \
- : ((t) == SDB_TYPE_BINARY) \
- ? "BINARY" \
- : "UNKNOWN")
+ (((t) == SDB_TYPE_INTEGER) ? "INTEGER" \
+ : ((t) == SDB_TYPE_DECIMAL) ? "DECIMAL" \
+ : ((t) == SDB_TYPE_STRING) ? "STRING" \
+ : ((t) == SDB_TYPE_DATETIME) ? "DATETIME" \
+ : ((t) == SDB_TYPE_BINARY) ? "BINARY" \
+ : ((t) == SDB_TYPE_REGEX) ? "REGEX" \
+ : ((t) == (SDB_TYPE_ARRAY | SDB_TYPE_INTEGER)) ? "[]INTEGER" \
+ : ((t) == (SDB_TYPE_ARRAY | SDB_TYPE_DECIMAL)) ? "[]DECIMAL" \
+ : ((t) == (SDB_TYPE_ARRAY | SDB_TYPE_STRING)) ? "[]STRING" \
+ : ((t) == (SDB_TYPE_ARRAY | SDB_TYPE_DATETIME)) ? "[]DATETIME" \
+ : ((t) == (SDB_TYPE_ARRAY | SDB_TYPE_BINARY)) ? "[]BINARY" \
+ : ((t) == (SDB_TYPE_ARRAY | SDB_TYPE_REGEX)) ? "[]REGEX" \
+ : "UNKNOWN")
+
+union sdb_datum;
+typedef union sdb_datum sdb_datum_t;
+
+union sdb_datum {
+ int64_t integer; /* SDB_TYPE_INTEGER */
+ double decimal; /* SDB_TYPE_DECIMAL */
+ char *string; /* SDB_TYPE_STRING */
+ sdb_time_t datetime; /* SDB_TYPE_DATETIME */
+ struct {
+ size_t length;
+ unsigned char *datum;
+ } binary; /* SDB_TYPE_BINARY */
+ struct {
+ char *raw;
+ regex_t regex;
+ } re; /* SDB_TYPE_REGEX */
+
+ struct {
+ size_t length;
+ void *values;
+ } array;
+};
/*
* sdb_data_t:
- * A datum retrieved from an arbitrary data source.
+ * An arbitrary value of a specified type.
*/
typedef struct {
- int type;
- union {
- int64_t integer; /* SDB_TYPE_INTEGER */
- double decimal; /* SDB_TYPE_DECIMAL */
- char *string; /* SDB_TYPE_STRING */
- sdb_time_t datetime; /* SDB_TYPE_DATETIME */
- struct {
- size_t length;
- unsigned char *datum;
- } binary; /* SDB_TYPE_BINARY */
- } data;
+ int type; /* type of the datum */
+ sdb_datum_t data;
} sdb_data_t;
+#define SDB_DATA_INIT { SDB_TYPE_NULL, { .integer = 0 } }
+
+extern const sdb_data_t SDB_DATA_NULL;
/*
* sdb_data_copy:
/*
* sdb_data_cmp:
* Compare two data points. A NULL datum is considered less than any non-NULL
- * datum. On data-type mismatch, the function always returns a negative value.
+ * datum. On data-type mismatch, the function always returns a non-zero value.
*
* Returns:
* - a value less than zero if d1 compares less than d2
int
sdb_data_cmp(const sdb_data_t *d1, const sdb_data_t *d2);
+/*
+ * sdb_data_strcmp:
+ * Compare the string values of two data points. A NULL datum is considered
+ * less than any non-NULL. This function works for arbitrary combination of
+ * data-types.
+ *
+ * Returns:
+ * - a value less than zero if d1 compares less than d2
+ * - zero if d1 compares equal to d2
+ * - a value greater than zero if d1 compares greater than d2
+ */
+int
+sdb_data_strcmp(const sdb_data_t *d1, const sdb_data_t *d2);
+
+/*
+ * sdb_data_isnull:
+ * Determine whether a datum is NULL. A datum is considered to be NULL if
+ * either datum is NULL or if the type is SDB_TYPE_NULL or if the string or
+ * binary datum is NULL.
+ */
+bool
+sdb_data_isnull(const sdb_data_t *datum);
+
+/*
+ * sdb_data_inarray:
+ * Determine whether a datum is included in an array based on the usual
+ * comparison function of the value's type. The element type of the array has
+ * to match the type of the value. The value may be another array. In that
+ * case, the element types have to match and the function returns true if all
+ * elements of the first array are included in the second where order does not
+ * matter.
+ */
+bool
+sdb_data_inarray(const sdb_data_t *value, const sdb_data_t *array);
+
+/*
+ * sdb_data_array_get:
+ * Get the i-th value stored in the specified array and store an alias in
+ * 'value'. Storing an alias means that the value points to the actual array
+ * element. Do *not* free the value after using it (i.e., don't use
+ * sdb_data_free_datum).
+ *
+ * Returns:
+ * - 0 on success
+ * - a negative value else
+ */
+int
+sdb_data_array_get(const sdb_data_t *array, size_t i, sdb_data_t *value);
+
+/*
+ * Operators supported by sdb_data_eval_expr.
+ */
+enum {
+ SDB_DATA_ADD = 1, /* addition */
+ SDB_DATA_SUB, /* substraction */
+ SDB_DATA_MUL, /* multiplication */
+ SDB_DATA_DIV, /* division */
+ SDB_DATA_MOD, /* modulo */
+ SDB_DATA_CONCAT, /* string / binary data concatenation */
+};
+
+#define SDB_DATA_OP_TO_STRING(op) \
+ (((op) == SDB_DATA_ADD) ? "+" \
+ : ((op) == SDB_DATA_SUB) ? "-" \
+ : ((op) == SDB_DATA_MUL) ? "*" \
+ : ((op) == SDB_DATA_DIV) ? "/" \
+ : ((op) == SDB_DATA_MOD) ? "%" \
+ : ((op) == SDB_DATA_CONCAT) ? "||" : "UNKNOWN")
+
+/*
+ * sdb_data_parse_op:
+ * Parse the string representation of an operator supported by
+ * sdb_data_expr_eval.
+ *
+ * Returns:
+ * - the ID of the operator
+ * - a negative value in case the operator does not exist
+ */
+int
+sdb_data_parse_op(const char *op);
+
+/*
+ * sdb_data_expr_eval:
+ * Evaluate a simple arithmetic expression on two data points. String and
+ * binary data only support concatenation and all other data types only
+ * support the other operators. The result may be allocated dynamically and
+ * has to be freed by the caller (using sdb_data_free_datum).
+ *
+ * If any of the data points is a NULL value, the result is also NULL.
+ *
+ * The data-types of d1 and d2 have to be the same, except for the following
+ * cases:
+ * - <integer> or <decimal> <mul> <datetime>
+ * - <datetime> <mul> or <div> or <mod> <integer> or <decimal>
+ *
+ * Returns:
+ * - 0 on success
+ * - a negative value else
+ */
+int
+sdb_data_expr_eval(int op, const sdb_data_t *d1, const sdb_data_t *d2,
+ sdb_data_t *res);
+
+/*
+ * sdb_data_expr_type:
+ * Determine the type of the expression when applying the specified operator
+ * to the specified types. Note that if an actual value is a typed NULL value
+ * (e.g. a NULL string value), the return value of this function does not
+ * match the return type of sdb_data_expr_eval.
+ *
+ * See the documentation of sdb_data_expr_eval() for a description of which
+ * operations are supported.
+ *
+ * Returns:
+ * - the type id on success
+ * - a negative value else
+ */
+int
+sdb_data_expr_type(int op, int type1, int type2);
+
/*
* sdb_data_strlen:
* Returns a (worst-case) estimate for the number of bytes required to format
@@ -157,7 +305,14 @@ sdb_data_format(const sdb_data_t *datum, char *buf, size_t buflen, int quoted);
* specified as (floating point) number of seconds since the epoch. For string
* and binary data, the input string is passed to the datum. The function does
* not allocate new memory for that purpose. Use sdb_data_copy() if you want
- * to do that.
+ * to do that. For regex data, the input string is copied to newly allocated
+ * memory and also compiled to a regex. Use sdb_data_free_datum() to free the
+ * dynamically allocated memory.
+ *
+ * The input string may be stored in 'data', that is, the function may be used
+ * to do an inline cast from a string to any other type. It is the callers
+ * responsibility to free the memory used by the string in case the target
+ * type does not keep a reference to it.
*
* Returns:
* - 0 on success
@@ -166,6 +321,17 @@ sdb_data_format(const sdb_data_t *datum, char *buf, size_t buflen, int quoted);
int
sdb_data_parse(char *str, int type, sdb_data_t *data);
+/*
+ * sdb_data_sizeof:
+ * Return the size of the data-type identified by the specified type.
+ *
+ * Returns:
+ * - the size of the data-type on success
+ * - 0 else
+ */
+size_t
+sdb_data_sizeof(int type);
+
#ifdef __cplusplus
} /* extern "C" */
#endif