* Scan json_view strings with SIMD and index large objects Speed up json_view's parser with SIMD scanning and a hash table for large objects. Long runs of string bytes are scanned 16 bytes at a time with NEON (AArch64, GCC and Clang) and SSE2 (x86-64), both baseline instruction sets. Keys keep 16 table checks before the vector loop, because their lengths repeat from record to record; string values get 8, because their lengths vary more. Non-ASCII text is validated 16 bytes at a time with simdjson's "lookup4" check (Keiser and Lemire, 2021), with NEON on AArch64 and, on x86-64, with SSSE3. SSSE3 is not part of baseline x86-64, so the check is compiled for SSSE3 with a function attribute and used only where CPUID reports it, which all x86-64 CPUs since about 2011 do; the answer is cached in a statically initialized atomic, so there is no guard of a local static and no global constructor. The same input is accepted either way. JSON_VIEW_NO_SIMD selects the portable code. On x86-64, string runs are now checked vector-first: one SSE2 compare from the first byte finds the end of most keys and short values, instead of a branch per byte for the first 8-16 bytes. AArch64 keeps the byte-wise steps, where a NEON mask costs more and the branches predict well. Entering an object or array no longer stalls: open() stores the parent's frame field by field instead of building it on the stack and reading it back with wider loads, which waited for the narrower stores to retire. Objects with 128 members or more get an open-addressing hash table built when the object closes, so operator[], at(), find(), contains(), count(), value(), and JSON pointers take constant time on average in such objects; of duplicate keys, the first is kept, as for the linear search. The idea comes from Boost.JSON. simdjson is credited in simd.hpp's SPDX block, the README, and license.md. Signed-off-by: Niels Lohmann <mail@nlohmann.me> * Fix GCC module build and MSVC C4127 in the json_view SIMD code GCC ignores the target attribute in modules, so the SSSE3 dispatch is disabled for the module interface (the check stays portable, SSE2 is kept). Make the 8-vs-16 byte unrolling condition in scan_string_run a preprocessor/template split to avoid a constant condition (C4127). Signed-off-by: Niels Lohmann <mail@nlohmann.me> * Remove an accidentally committed GCC module cache Signed-off-by: Niels Lohmann <mail@nlohmann.me> * Bound the probe length of the large-object hash index The key hash is not seeded, so keys chosen to collide made building the table quadratic (20,000 colliding keys took 470 ms to parse). A key may now sit at most 64 slots from its home slot; if a key would sit further away, the table is dropped and the object is searched linearly. Lookups stop after the same distance. Signed-off-by: Niels Lohmann <mail@nlohmann.me> * Release the spare capacity of the hash index in shrink_to_fit shrink_to_fit() now trims the tables of large objects like the node array and the decoded strings, and the list of large objects is released as soon as the tables are built. Signed-off-by: Niels Lohmann <mail@nlohmann.me> * Use the portable string scan on non-x86 targets that define __SSE2__ WebAssembly with -msse2 defines __SSE2__, but has no <cpuid.h> or x86 intrinsics headers. Signed-off-by: Niels Lohmann <mail@nlohmann.me> * Document the hash index number in node::extra Signed-off-by: Niels Lohmann <mail@nlohmann.me> * Keep the last of duplicate keys in the object hash index Lookups in objects with 128 members or more now return the last member of a repeated key, like the linear search of smaller objects and like materialize() and basic_json::parse(). build_object_index let the first occurrence win, so the same text gave different results depending on the size of the object. Signed-off-by: Niels Lohmann <mail@nlohmann.me> * Keep the first of duplicate keys in the object hash index again Lookups in objects with 128 members or more return the first member of a repeated key again, like the linear search of smaller objects. This undoes the code change of a69542046; its test now expects the first member from lookups (with and without a table) and the last value from materialize() and basic_json::parse(). Signed-off-by: Niels Lohmann <mail@nlohmann.me> * Fix clang-tidy 22 findings in the json_view headers Silence bugprone-casting-through-void for the SSE loads (a reinterpret_cast would trip -Wcast-align=strict), use auto for a cast initialiser, add parentheses to a mixed expression, and name bugprone-std-namespace-modification in the NOLINTs of the tuple_size/tuple_element specialisations. Signed-off-by: Niels Lohmann <mail@nlohmann.me> * Fix clang-tidy 22 findings in unit-json_view.cpp and unit-json_view_builder.cpp Signed-off-by: Niels Lohmann <mail@nlohmann.me> * Avoid GCC useless-cast and strict-overflow warnings in unit-json_view.cpp Signed-off-by: Niels Lohmann <mail@nlohmann.me> * Index strings with size_t in the colliding-keys test (MSVC C4244 on Win32) Signed-off-by: Niels Lohmann <mail@nlohmann.me> * Make json_view.hpp pass include-what-you-use ci_single_binaries runs IWYU with --error on every header. For json_view.hpp it suggested adding <array> (std::array is used), the headers that json.hpp already provides (abi_config, abi_macros, input_adapters, json_pointer, cpp_future, string_concat, value_t, json_fwd), and <version> for std::nullptr_t, and removing <cstddef>. - include <array> - keep <cstddef> (nullptr_t, size_t; IWYU attributes them to <version> and <cstring>) - tell IWYU not to suggest the headers that json.hpp provides: the amalgamated json_view.hpp only includes json.hpp, so including them here would duplicate their definitions Signed-off-by: Niels Lohmann <mail@nlohmann.me> * Move the dump, comparison, and large-object tests to unit-json_view_dump.cpp The dump and comparison tests moved to their own file in the dump pull request; the hash index tests of this branch join them there. Signed-off-by: Niels Lohmann <mail@nlohmann.me> * Regenerate the amalgamated headers Signed-off-by: Niels Lohmann <mail@nlohmann.me> * Follow the renames from the review of the view builder The review of #5620 renamed document_data's tape/tape_size/tape_cap to nodes/node_count/node_capacity and made the decode arena a std::vector<char>; adapt the code of this change to it. Signed-off-by: Niels Lohmann <mail@nlohmann.me> * Rename slots in the object index for Qt compatibility Qt defines slots as a macro, so a Qt header included before json_view.hpp broke the declarations of the local variable (see #5865). The test variable is renamed as well. Signed-off-by: Niels Lohmann <mail@nlohmann.me> * Address the review of the json_view object index and SIMD scan From gregmarr's review of #5628: - Docs and comments say that the hash index is one of the object's keys. - object_index.hpp: conditional_static_cast for the hash, "slot" for the table position, no flag for a degenerate table, and a static_assert that the smallest indexed object is larger than the probe window. - scan.hpp unrolls with NLOHMANN_VIEW_REPEAT_FIRST_8/SECOND_8. Signed-off-by: Niels Lohmann <mail@nlohmann.me> * Regenerate the amalgamated headers Signed-off-by: Niels Lohmann <mail@nlohmann.me> * Encode JSON_VIEW_NO_SIMD and JSON_VIEW_USE_SSSE3 in the ABI namespace Both macros select between definitions of the same inline functions, so translation units that disagree on them violated the ODR. Like the other ABI-affecting macros, they now add a tag to the inline namespace (_vns and _vssse3), and the docs replace the "Define consistently" warning with the usual ABI compatibility note. From gregmarr's review of #5628. nlohmann_json.natvis covers twice as many tags now, so it lists four times as many namespaces (14.7 MB instead of 3.4 MB). Signed-off-by: Niels Lohmann <mail@nlohmann.me> * Regenerate the amalgamated headers Signed-off-by: Niels Lohmann <mail@nlohmann.me> * Do not include type_traits.hpp directly in the object index json.hpp, which object_index.hpp already includes, provides it; included directly, the json_view amalgamation (which only treats nlohmann/json.hpp as external) copied the json internals into json_view.hpp a second time, so the fuzzers, which include both single headers, failed with redefinitions (make_void, integer_sequence, priority_tag, ...). Signed-off-by: Niels Lohmann <mail@nlohmann.me> * Compare keys with key_equals() in the object hash index As the linear lookups do since the second review round of #5622. Signed-off-by: Niels Lohmann <mail@nlohmann.me> * Regenerate the amalgamated headers Signed-off-by: Niels Lohmann <mail@nlohmann.me> --------- Signed-off-by: Niels Lohmann <mail@nlohmann.me>
9.3 KiB
nlohmann::basic_json_view::operator[]
// (1)
basic_json_view operator[](string_view_t key) const;
basic_json_view operator[](const char* key) const;
basic_json_view operator[](const string_t& key) const;
// (2)
basic_json_view operator[](size_type idx) const;
template<typename IntegerType>
basic_json_view operator[](IntegerType idx) const;
// (3)
basic_json_view operator[](const json_pointer& ptr) const;
- Returns the value of the object member with key
key-- the first one, should the key occur more than once (see the Notes below) -- or a discarded view if there is no such member. - Returns the array element at index
idx, or a discarded view ifidxis out of range. The template accepts every integer type except#!cpp booland#!cpp std::size_t(#!cpp int,#!cpp unsigned,#!cpp long,#!cpp std::int64_t, ...) and forwards to thesize_typeoverload, so that an integer argument is not ambiguous between that overload and 1; a negativeidxis out of range. - Returns the value a JSON pointer
ptrrefers to, starting at this value, or a discarded view wherever resolving it further is not possible without inserting into or extending the document (see Return value and Exceptions below).
Parameters
key(in)- object key of the element to access
idx(in)- index of the element to access
ptr(in)- JSON pointer to the element to access
Return value
- the value of the first member with key
key, or a discarded view if no member has this key (or if this view is discarded) - the element at index
idx, or a discarded view if#!cpp idx >= size()oridxis negative (or if this view is discarded) - the value
ptrresolves to, starting at this value, or a discarded view (also if this view is discarded) for exactly the reference tokens where the const overload ofBasicJsonType::operator[]invokes undefined behavior for the same pointer and the same document: an object member that does not exist, or an array index that is out of range
Exception safety
Strong guarantee: there are no changes to the view or the document it refers to.
Exceptions
- Throws
type_error.305if the value is not an object and not discarded -- the same exception, with the same message, that the const overload ofBasicJsonType::operator[]throws for a string argument on a non-object value. - Throws
type_error.305if the value is not an array and not discarded -- the same exception, with the same message, that the const overload ofBasicJsonType::operator[]throws for a numeric argument on a non-array value. - Throws the same exceptions, with the same messages, that the const overload of
BasicJsonType::operator[]throws for the same pointer and the same document:out_of_range.402if a reference token is#!cpp "-"at an array.out_of_range.404if a reference token cannot be resolved because it is used on a primitive value.parse_error.106if an array index inptrbegins with#!cpp '0'.parse_error.109if an array index inptris not a number.
None of these exceptions carry a JSON_DIAGNOSTICS path: the view has no
BasicJsonType value to point at, so the exception is created without one, even if BasicJsonType was built with
JSON_DIAGNOSTICS enabled.
Complexity
- Linear in the number of members: as for
ordered_json, members are compared one after another, in document order, stopping at the first match. Each comparison first checks the key's length -- already known from the index, without reading the key bytes -- before comparing its content, so a key of a different length thankeyis rejected without touching the source text. Objects with 128 or more members get a hash index of their keys while parsing, so that a lookup in them takes constant time on average. - Linear in
idx: elements are skipped one at a time from the first one, since they are not a fixed size in the index (unlikeBasicJsonType's array, which is random-access). - Linear in the number of reference tokens of
ptrand, for each token, in the number of members of the object at that level (as 1.) or the index into the array (as 2.).
Notes
Unlike BasicJsonType::operator[], which is undefined behavior (guarded by a
runtime assertion) for a missing key on a const value, this operator returns a
safe, testable result for a missing key or an index out of range: a discarded view, which is
tested with is_discarded.
There is also no non-const overload that inserts a missing key or extends an array -- a view never modifies the
document.
!!! info "Chained access"
`#!cpp operator[]` on a [discarded](is_discarded.md) view returns a discarded view and does not throw, so a chain
like `#!cpp v["a"]["b"][0]` is safe even if `"a"` or `"b"` is missing: the first missing step makes the whole
result discarded, which is tested once at the end. Type errors on values that are *not* discarded still throw: a
key on an array or a primitive, or an index on an object or a primitive, is `type_error.305` as for
`BasicJsonType`. [`at`](at.md) still throws for a discarded view, as it does for a missing key.
!!! warning "Duplicate keys: the first member wins"
If the source text has an object with a duplicate key, `#!cpp operator[]` (and [`at`](at.md), [`find`](find.md),
[`contains`](contains.md), [`value`](value.md), and JSON pointer resolution) all resolve to
the *first* member with that key, because a lookup can stop as soon as it finds a match. This is different from
[`materialize()`](materialize.md) (and [`BasicJsonType::parse()`](../basic_json/parse.md)), which replay every
member in order and so keep the *last* value for a repeated key, so `#!cpp v["a"]` and
`#!cpp v.materialize()["a"]` can differ. To get the value `parse()` would give, use
[`materialize()`](materialize.md) or iterate the members with [`items()`](items.md) and keep the last match.
The key hash index of a larger object (128 members or more) leads to the first member of a key as well.
[`begin()`](begin.md)/[`end()`](end.md) and [`items()`](items.md) iterate over *all* members, including
duplicates, in document order. See [Duplicate keys](../../features/json_view.md#duplicate-keys) and
[`size()`](size.md#notes).
!!! info "JSON pointer resolution"
Overload 3 walks `ptr` one reference token at a time, starting at this value, the same way [`at`](at.md) and
[`contains`](contains.md) do. It only ever returns a discarded view where the **const** overload of
`BasicJsonType::operator[]` would be undefined behavior for the same pointer -- a missing object member or an
out-of-range array index -- and still throws for every other way `ptr` can fail to resolve. See
[`at`](at.md#exceptions) for the checked version, which throws in every case instead, and
[`contains`](contains.md) for a version that never throws.
Examples
??? example "Example: (1)/(2) access specified element"
The example below reads a couple of fields out of a batch of user records without ever materializing a full
`BasicJsonType` value for the batch. `operator[]` is used both to look up an optional object member and to index
into an array -- in both cases, a missing value comes back as a discarded view that can be tested with a plain
`#!cpp if`, instead of relying on undefined behavior.
```cpp
--8<-- "examples/basic_json_view__operator[].cpp"
```
Output:
```json
--8<-- "examples/basic_json_view__operator[].output"
```
??? example "Example: (3) access specified element via JSON pointer"
The example below reaches straight into one deeply nested field of a large document with a single JSON pointer,
without ever building a tree for the rest of it, and shows the discarded-view and throwing outcomes of a pointer
that cannot be fully resolved.
```cpp
--8<-- "examples/basic_json_view__operator[]_json_pointer.cpp"
```
Output:
```json
--8<-- "examples/basic_json_view__operator[]_json_pointer.output"
```
See also
- at - access specified element with bounds checking (throws instead of returning a discarded view)
- front, back - access the first or last element
- find, contains - look up a member without throwing
BasicJsonType::operator[]- the corresponding function ofbasic_jsonjson_pointer- JSON pointer type used by overload 3
Version history
- Added in version 3.13.0.