ExtensionFunctionsText.hpp File Reference

#include <cstring>
#include "Shared/toString.h"
#include "heavydbTypes.h"

Include dependency graph for ExtensionFunctionsText.hpp:

This graph shows which files directly or indirectly include this file:

Go to the source code of this file.

Functions
std::vector< std::string >	__strtok_to_array (const std::string &text, const std::string &delimiters)
EXTENSION_NOINLINE Array < TextEncodingDict >	strtok_to_array (RowFunctionManager &mgr, TextEncodingNone &text, TextEncodingNone &delimiters)
EXTENSION_NOINLINE Array < TextEncodingDict >	strtok_to_array__1 (RowFunctionManager &mgr, TextEncodingDict text, TextEncodingNone &delimiters)

Function Documentation

std::vector<std::string> __strtok_to_array	(	const std::string &	text,
		const std::string &	delimiters
	)

Definition at line 24 of file ExtensionFunctionsText.hpp.

Referenced by strtok_to_array(), and strtok_to_array__1().

                                                                        {
  std::vector<std::string> vec;

  char* str = const_cast<char*>(text.c_str());
  const char* del = delimiters.c_str();

  char* substr = strtok(str, del);
  while (substr != NULL) {
    std::string s(substr);
    vec.emplace_back(s);
    substr = strtok(NULL, del);
  }

  return vec;
}

Here is the caller graph for this function:

EXTENSION_NOINLINE Array<TextEncodingDict> strtok_to_array	(	RowFunctionManager &	mgr,
		TextEncodingNone &	text,
		TextEncodingNone &	delimiters
	)

Definition at line 42 of file ExtensionFunctionsText.hpp.

References __strtok_to_array(), RowFunctionManager::getOrAddTransient(), TextEncodingNone::getString(), TextEncodingNone::isNull(), TRANSIENT_DICT_DB_ID, and TRANSIENT_DICT_ID.

                                                                      {
  /*
    Rules
    -----
    * If either parameters is NULL => a NULL is returned
    * An empty array is returned if tokenization produces no tokens

    Note
    ----
    <delimiters> argument is optional on snowflake but HeavyDB dont' support
    default values on UDFs at the moment. See:
    https://github.com/heavyai/heavydb-internal/pull/6651

    Examples
    --------
    > select strtok_to_array('a.b.c', '.');
    {a, b, c}

    > select strtok_to_array('user@gmail.com', '.@')
    {user, gmail, com}

    > select strtok_to_array('', '.')
    NULL

    > select strtok_to_array('a.b.c', '')
    NULL
  */

  if (text.isNull() || delimiters.isNull()) {
    return Array<TextEncodingDict>(0, true);
  }

  const auto& vec = __strtok_to_array(text.getString(), delimiters.getString());
  Array<TextEncodingDict> out_arr(vec.size());
  for (size_t i = 0; i < vec.size(); ++i) {
    out_arr[i] = mgr.getOrAddTransient(TRANSIENT_DICT_DB_ID, TRANSIENT_DICT_ID, vec[i]);
  }
  return out_arr;
}

Here is the call graph for this function:

EXTENSION_NOINLINE Array<TextEncodingDict> strtok_to_array__1	(	RowFunctionManager &	mgr,
		TextEncodingDict	text,
		TextEncodingNone &	delimiters
	)

Definition at line 85 of file ExtensionFunctionsText.hpp.

References __strtok_to_array(), GET_DICT_DB_ID, GET_DICT_ID, RowFunctionManager::getOrAddTransient(), TextEncodingNone::getString(), RowFunctionManager::getString(), TextEncodingDict::isNull(), TextEncodingNone::isNull(), TRANSIENT_DICT_DB_ID, and TRANSIENT_DICT_ID.

                                                                         {
  if (text.isNull() || delimiters.isNull()) {
    return Array<TextEncodingDict>(0, true);
  }

  std::string str = mgr.getString(GET_DICT_DB_ID(mgr, 0), GET_DICT_ID(mgr, 0), text);
  const auto& vec = __strtok_to_array(str, delimiters.getString());
  Array<TextEncodingDict> out_arr(vec.size());
  for (size_t i = 0; i < vec.size(); ++i) {
    out_arr[i] = mgr.getOrAddTransient(TRANSIENT_DICT_DB_ID, TRANSIENT_DICT_ID, vec[i]);
  }
  return out_arr;
}

Here is the call graph for this function: