mirrors/scylladb
1
0
Fork 0
mirror of https://github.com/scylladb/scylladb.git synced 2026-06-02 13:06:57 +00:00
Code Issues Packages Projects Releases Wiki Activity

tests/random_schema: generate_random_mutations(): allow customizing generated data

Browse Source 
Allow callers to specify the number of partitions generated, as well as
the number of clustering rows and range tombstones generated per
partition.
This commit is contained in:
Botond Dénes
2019-07-05 10:17:24 +03:00
parent d2930ffa53
commit cf135c6257
3 changed files with 44 additions and 20 deletions
Download Patch File Download Diff File Expand all files Collapse all files

15
tests/mutation_writer_test.cc
View File

@@ -241,7 +241,20 @@ SEASTAR_THREAD_TEST_CASE(test_timestamp_based_splitting_mutation_writer) {
tlog.info("Random schema:\n{}", random_schema.cql());
auto muts = tests::generate_random_mutations(random_schema);
auto ts_gen = [&, underlying = tests::default_timestamp_generator()] (std::mt19937& engine,
tests::timestamp_destination ts_dest, api::timestamp_type min_timestamp) -> api::timestamp_type {
if (ts_dest == tests::timestamp_destination::partition_tombstone ||
ts_dest == tests::timestamp_destination::row_marker ||
ts_dest == tests::timestamp_destination::row_tombstone ||
ts_dest == tests::timestamp_destination::collection_tombstone) {
if (tests::random::get_int<int>(0, 10, engine)) {
return api::missing_timestamp;
}
}
return underlying(engine, ts_dest, min_timestamp);
};
auto muts = tests::generate_random_mutations(random_schema, ts_gen);
auto classify_fn = [] (api::timestamp_type ts) {
return int64_t(ts % 2);

37
tests/random_schema.cc
View File

@@ -1007,29 +1007,30 @@ void random_schema::delete_range(
md.add_range_tombstone(std::move(range), tombstone{ts_gen(engine, timestamp_destination::range_tombstone, api::min_timestamp), {}});
}
std::vector<mutation> generate_random_mutations(tests::random_schema& random_schema) {
auto ts_gen = [&, underlying = tests::default_timestamp_generator()] (std::mt19937& engine,
tests::timestamp_destination ts_dest, api::timestamp_type min_timestamp) -> api::timestamp_type {
if (ts_dest == tests::timestamp_destination::partition_tombstone ||
ts_dest == tests::timestamp_destination::row_marker ||
ts_dest == tests::timestamp_destination::row_tombstone ||
ts_dest == tests::timestamp_destination::collection_tombstone) {
if (tests::random::get_int<int>(0, 10, engine)) {
return api::missing_timestamp;
}
}
return underlying(engine, ts_dest, min_timestamp);
};
std::vector<mutation> generate_random_mutations(
tests::random_schema& random_schema,
timestamp_generator ts_gen,
std::uniform_int_distribution<size_t> partition_count_dist,
std::uniform_int_distribution<size_t> clustering_row_count_dist,
std::uniform_int_distribution<size_t> range_tombstone_count_dist) {
auto engine = std::mt19937(tests::random::get_int<uint32_t>());
const auto schema_has_clustering_columns = random_schema.schema()->clustering_key_size() > 0;
const auto partition_count = partition_count_dist(engine);
auto muts = std::vector<mutation>{};
auto ckeys = random_schema.make_ckeys(100);
for (uint32_t pk = 0; pk < 10; ++pk) {
for (uint32_t pk = 0; pk < partition_count; ++pk) {
auto mut = random_schema.new_mutation(pk);
for (uint32_t ck = 0; ck < 100; ++ck) {
random_schema.add_static_row(engine, mut, ts_gen);
if (!schema_has_clustering_columns) {
muts.emplace_back(mut.build(random_schema.schema()));
continue;
}
const auto clustering_row_count = clustering_row_count_dist(engine);
auto ckeys = random_schema.make_ckeys(clustering_row_count);
for (uint32_t ck = 0; ck < clustering_row_count; ++ck) {
random_schema.add_row(engine, mut, ckeys[ck], ts_gen);
}
random_schema.add_static_row(engine, mut, ts_gen);
for (size_t i = 0; i < 4; ++i) {
const auto a = tests::random::get_int<size_t>(0, ckeys.size() - 1, engine);

12
tests/random_schema.hh
View File

@@ -220,6 +220,16 @@ public:
timestamp_generator ts_gen = default_timestamp_generator());
};
std::vector<mutation> generate_random_mutations(tests::random_schema& random_schema);
/// Generate random mutations using the random schema.
///
/// `clustering_row_count_dist` and `range_tombstone_count_dist` will be used to
/// generate the respective counts for *each* partition. These params are
/// ignored if the schema has no clustering columns.
std::vector<mutation> generate_random_mutations(
tests::random_schema& random_schema,
timestamp_generator ts_gen = default_timestamp_generator(),
std::uniform_int_distribution<size_t> partition_count_dist = std::uniform_int_distribution<size_t>(8, 16),
std::uniform_int_distribution<size_t> clustering_row_count_dist = std::uniform_int_distribution<size_t>(16, 128),
std::uniform_int_distribution<size_t> range_tombstone_count_dist = std::uniform_int_distribution<size_t>(4, 16));
} // namespace tests