jasder
/
simdjson
1
0
Fork 0
Code Issues Pull Requests Releases 2 Wiki Activity

Descend into fields at the value position, not the key

This commit is contained in:
John Keiser 2021-01-28 22:33:19 -08:00
parent 9934f65987
commit e4626d233c
5 changed files with 20 additions and 27 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
include/simdjson/generic/ondemand/json_iterator-inl.h
View File

@ -154,13 +154,10 @@ simdjson_really_inline void json_iterator::ascend_to(depth_t parent_depth) noexc
} }
simdjson_really_inline void json_iterator::descend_to(depth_t child_depth) noexcept { simdjson_really_inline void json_iterator::descend_to(depth_t child_depth) noexcept {
descend_to(child_depth, token.index);
}
simdjson_really_inline void json_iterator::descend_to(depth_t child_depth, token_position start_position) noexcept {
SIMDJSON_ASSUME(child_depth >= 1 && child_depth < INT32_MAX); SIMDJSON_ASSUME(child_depth >= 1 && child_depth < INT32_MAX);
SIMDJSON_ASSUME(_depth == child_depth - 1); SIMDJSON_ASSUME(_depth == child_depth - 1);
_depth = child_depth; _depth = child_depth;
parser->start_positions[_depth] = start_position; parser->start_positions[_depth] = token.index;
} }
simdjson_really_inline depth_t json_iterator::depth() const noexcept { simdjson_really_inline depth_t json_iterator::depth() const noexcept {
@ -181,11 +178,12 @@ simdjson_really_inline error_code json_iterator::report_error(error_code _error,
simdjson_really_inline token_position json_iterator::position() const noexcept { simdjson_really_inline token_position json_iterator::position() const noexcept {
return token.position(); return token.position();
} }
simdjson_really_inline void json_iterator::set_position(token_position target_checkpoint) noexcept { simdjson_really_inline void json_iterator::reenter_child(token_position position, depth_t child_depth) noexcept {
token.set_position(target_checkpoint); SIMDJSON_ASSUME(child_depth >= 1 && child_depth < INT32_MAX);
} SIMDJSON_ASSUME(_depth == child_depth - 1);
simdjson_really_inline token_position json_iterator::nested_start_position(depth_t depth) const noexcept { SIMDJSON_ASSUME(position >= parser->start_positions[child_depth]);
return parser->start_positions[depth]; token.set_position(position);
_depth = child_depth;
} }

5
include/simdjson/generic/ondemand/json_iterator.h
View File

@ -148,7 +148,7 @@ public:
* @param child_depth the expected child depth. * @param child_depth the expected child depth.
*/ */
simdjson_really_inline void descend_to(depth_t parent_depth) noexcept; simdjson_really_inline void descend_to(depth_t parent_depth) noexcept;
simdjson_really_inline void descend_to(depth_t parent_depth, token_position start_position) noexcept; simdjson_really_inline void descend_to(depth_t parent_depth, int32_t delta) noexcept;
/** /**
* Get current depth. * Get current depth.
@ -180,8 +180,7 @@ public:
template<int N> simdjson_warn_unused simdjson_really_inline bool advance_to_buffer(uint8_t (&tmpbuf)[N]) noexcept; template<int N> simdjson_warn_unused simdjson_really_inline bool advance_to_buffer(uint8_t (&tmpbuf)[N]) noexcept;
simdjson_really_inline token_position position() const noexcept; simdjson_really_inline token_position position() const noexcept;
simdjson_really_inline void set_position(token_position target_checkpoint) noexcept; simdjson_really_inline void reenter_child(token_position position, depth_t child_depth) noexcept;
simdjson_really_inline token_position nested_start_position(depth_t depth) const noexcept;
protected: protected:
simdjson_really_inline json_iterator(const uint8_t *buf, ondemand::parser *parser) noexcept; simdjson_really_inline json_iterator(const uint8_t *buf, ondemand::parser *parser) noexcept;

1
include/simdjson/generic/ondemand/parser.h
View File

@ -123,6 +123,7 @@ private:
simdjson_warn_unused error_code allocate(size_t capacity, size_t max_depth=DEFAULT_MAX_DEPTH) noexcept; simdjson_warn_unused error_code allocate(size_t capacity, size_t max_depth=DEFAULT_MAX_DEPTH) noexcept;
friend class json_iterator; friend class json_iterator;
friend class value_iterator;
}; };
} // namespace ondemand } // namespace ondemand

18
include/simdjson/generic/ondemand/value_iterator-inl.h
View File

@ -24,7 +24,6 @@ simdjson_warn_unused simdjson_really_inline bool value_iterator::started_object(
_json_iter->ascend_to(depth()-1); _json_iter->ascend_to(depth()-1);
return false; return false;
} }
_json_iter->descend_to(depth()+1); //, _start_position+3); // skip {"key":
logger::log_start_value(*_json_iter, "object"); logger::log_start_value(*_json_iter, "object");
return true; return true;
} }
@ -38,7 +37,6 @@ simdjson_warn_unused simdjson_really_inline simdjson_result<bool> value_iterator
_json_iter->ascend_to(depth()-1); _json_iter->ascend_to(depth()-1);
return false; return false;
case ',': case ',':
_json_iter->descend_to(depth()+1); //, _json_iter->token.index+2); // index+2 skips "key":
return true; return true;
default: default:
return _json_iter->report_error(TAPE_ERROR, "Missing comma between object fields"); return _json_iter->report_error(TAPE_ERROR, "Missing comma between object fields");
@ -94,9 +92,9 @@ simdjson_warn_unused simdjson_really_inline simdjson_result<bool> value_iterator
// ``` // ```
// //
} else { } else {
// if (_json_iter->nested_start_position(depth()) != _start_position+1) { return OUT_OF_ORDER_ITERATION; }
if ((error = skip_child() )) { abandon(); return error; } if ((error = skip_child() )) { abandon(); return error; }
if ((error = has_next_field().get(has_value) )) { abandon(); return error; } if ((error = has_next_field().get(has_value) )) { abandon(); return error; }
// if (_json_iter->parser->start_positions[_depth] != _start_position) { return OUT_OF_ORDER_ITERATION; }
} }
while (has_value) { while (has_value) {
// Get the key and colon, stopping at the value. // Get the key and colon, stopping at the value.
@ -169,10 +167,10 @@ simdjson_warn_unused simdjson_really_inline simdjson_result<bool> value_iterator
// ``` // ```
// //
} else { } else {
// if (_json_iter->nested_start_position(depth()) != _start_position) { return OUT_OF_ORDER_ITERATION; }
// Finish the previous value and see if , or } is next // Finish the previous value and see if , or } is next
if ((error = skip_child() )) { abandon(); return error; } if ((error = skip_child() )) { abandon(); return error; }
if ((error = has_next_field().get(has_value) )) { abandon(); return error; } if ((error = has_next_field().get(has_value) )) { abandon(); return error; }
// if (_json_iter->parser->start_positions[_depth] != _start_position) { return OUT_OF_ORDER_ITERATION; }
} }
// After initial processing, we will be in one of two states: // After initial processing, we will be in one of two states:
@ -195,7 +193,7 @@ simdjson_warn_unused simdjson_really_inline simdjson_result<bool> value_iterator
// Next, we find a match starting from the current position. // Next, we find a match starting from the current position.
while (has_value) { while (has_value) {
SIMDJSON_ASSUME( _json_iter->_depth == _depth + 1 ); // We must be at the start of a field SIMDJSON_ASSUME( _json_iter->_depth == _depth ); // We must be at the start of a field
// Get the key and colon, stopping at the value. // Get the key and colon, stopping at the value.
raw_json_string actual_key; raw_json_string actual_key;
@ -218,13 +216,12 @@ simdjson_warn_unused simdjson_really_inline simdjson_result<bool> value_iterator
// beginning of the object. // beginning of the object.
// (We have already run through the object before, so we've already validated its structure. We // (We have already run through the object before, so we've already validated its structure. We
// don't check errors in this bit.) // don't check errors in this bit.)
_json_iter->set_position(_start_position + 1); _json_iter->reenter_child(_start_position + 1, _depth);
_json_iter->descend_to(depth()); // , _start_position);
has_value = started_object(); has_value = started_object();
while (_json_iter->position() < search_start) { while (_json_iter->position() < search_start) {
SIMDJSON_ASSUME(has_value); // we should reach search_start before ever reaching the end of the object SIMDJSON_ASSUME(has_value); // we should reach search_start before ever reaching the end of the object
SIMDJSON_ASSUME( _json_iter->_depth == _depth + 1 ); // We must be at the start of a field SIMDJSON_ASSUME( _json_iter->_depth == _depth ); // We must be at the start of a field
// Get the key and colon, stopping at the value. // Get the key and colon, stopping at the value.
raw_json_string actual_key; raw_json_string actual_key;
@ -248,7 +245,7 @@ simdjson_warn_unused simdjson_really_inline simdjson_result<bool> value_iterator
} }
simdjson_warn_unused simdjson_really_inline simdjson_result<raw_json_string> value_iterator::field_key() noexcept { simdjson_warn_unused simdjson_really_inline simdjson_result<raw_json_string> value_iterator::field_key() noexcept {
assert_at_child(); assert_at_next();
const uint8_t *key = _json_iter->advance(); const uint8_t *key = _json_iter->advance();
if (*(key++) != '"') { return _json_iter->report_error(TAPE_ERROR, "Object key is not a string"); } if (*(key++) != '"') { return _json_iter->report_error(TAPE_ERROR, "Object key is not a string"); }
@ -256,9 +253,10 @@ simdjson_warn_unused simdjson_really_inline simdjson_result<raw_json_string> val
} }
simdjson_warn_unused simdjson_really_inline error_code value_iterator::field_value() noexcept { simdjson_warn_unused simdjson_really_inline error_code value_iterator::field_value() noexcept {
assert_at_child(); assert_at_next();
if (*_json_iter->advance() != ':') { return _json_iter->report_error(TAPE_ERROR, "Missing colon in object field"); } if (*_json_iter->advance() != ':') { return _json_iter->report_error(TAPE_ERROR, "Missing colon in object field"); }
_json_iter->descend_to(depth()+1);
return SUCCESS; return SUCCESS;
} }

7
tests/ondemand/ondemand_object_tests.cpp
View File

@ -508,11 +508,8 @@ namespace object_tests {
SUBTEST("ondemand::value", test_ondemand_doc(json, [&](auto doc_result) { SUBTEST("ondemand::value", test_ondemand_doc(json, [&](auto doc_result) {
ondemand::value object; ondemand::value object;
ASSERT_SUCCESS( doc_result["outer"].get(object) ); ASSERT_SUCCESS( doc_result["outer"].get(object) );
uint64_t v; ASSERT_EQUAL( object["a"].get_uint64().value_unsafe(), 1 );
ASSERT_SUCCESS( object["a"].get(v) ); ASSERT_EQUAL( object["b"].get_uint64().value_unsafe(), 2 );
ASSERT_EQUAL( v, 1 );
ASSERT_SUCCESS( object["b"].get(v) );
ASSERT_EQUAL( v, 2 );
ASSERT_EQUAL( object["c/d"].get_uint64().value_unsafe(), 3 ); ASSERT_EQUAL( object["c/d"].get_uint64().value_unsafe(), 3 );
ASSERT_EQUAL( object["a"].get_uint64().value_unsafe(), 1 ); ASSERT_EQUAL( object["a"].get_uint64().value_unsafe(), 1 );