Previous set of QueryProcessingStage does not allow to do this.
But after WithMergeableStateAfterAggregation had been introduced the
following queries can be optimized too under
optimize_distributed_group_by_sharding_key:
- GROUP BY sharding_key LIMIT
- GROUP BY sharding_key LIMIT BY
- GROUP BY sharding_key ORDER BY

And right now it is still not supports:
- WITH TOTALS (looks like it can be supported)
- WITH ROLLUP (looks like it can be supported)
- WITH CUBE
- SETTINGS extremes=1 (looks like it can be supported)
But will be implemented separatelly.

vX: fixes
v2: fix WITH *
v3: fix extremes
v4: fix LIMIT OFFSET (and make a little bit cleaner)
v5: fix HAVING
v6: fix ORDER BY
v7: rebase against 20.7
v8: move out WithMergeableStateAfterAggregation
v9: add optimize_distributed_group_by_sharding_key into test names

(cherry picked from commit 81481f26b5d1136daddc3265a90f52e4ef48a1c4)

(cherry picked from commit 5daee1cae199816a4f3e8fbc466a8974999b772e)
(cherry picked from commit c3dfb8e83844ae73c1a9879fbc9b0d9353bfbb6d)

Example of such queries are distributed queries, which creates local
InterpreterSelectQuery, which will have it's own QueryPlan but returns
Pipes that has that IQueryPlanStep attached.

After EXPLAIN PIPELINE graph=1 tries to use them, and will get SIGSEGV.

- TSAN:

<details>

```
==2782113==ERROR: AddressSanitizer: heap-use-after-free on address 0x6120000223c0 at pc 0x00002b8f3f3e bp 0x7fff18cfbff0 sp 0x7fff18cfbfe8
READ of size 8 at 0x6120000223c0 thread T22 (TCPHandler)
    #0 0x2b8f3f3d in DB::printPipelineCompact(std::__1::vector<std::__1::shared_ptr<DB::IProcessor>, std::__1::allocator<std::__1::shared_ptr<DB::IProcessor> > > const&, DB::WriteBuffer&, bool) /build/obj-x86_64-linux-gnu/../src/Processors/printPipeline.cpp:116:53
    #1 0x29ee698c in DB::InterpreterExplainQuery::executeImpl() /build/obj-x86_64-linux-gnu/../src/Interpreters/InterpreterExplainQuery.cpp:275:17
    #2 0x29ee2e40 in DB::InterpreterExplainQuery::execute() /build/obj-x86_64-linux-gnu/../src/Interpreters/InterpreterExplainQuery.cpp:73:14
    #3 0x2a7b44a2 in DB::executeQueryImpl(char const*, char const*, DB::Context&, bool, DB::QueryProcessingStage::Enum, bool, DB::ReadBuffer*) /build/obj-x86_64-linux-gnu/../src/Interpreters/executeQuery.cpp:389:28
    #4 0x2a7b1cb3 in DB::executeQuery(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, DB::Context&, bool, DB::QueryProcessingStage::Enum, bool) /build/obj-x86_64-linux-gnu/../src/Interpreters/executeQuery.cpp:675:30
    #5 0x2b7993b2 in DB::TCPHandler::runImpl() /build/obj-x86_64-linux-gnu/../src/Server/TCPHandler.cpp:253:24
    #6 0x2b7b649a in DB::TCPHandler::run() /build/obj-x86_64-linux-gnu/../src/Server/TCPHandler.cpp:1217:9
    #7 0x31d9c57e in Poco::Net::TCPServerConnection::start() /build/obj-x86_64-linux-gnu/../contrib/poco/Net/src/TCPServerConnection.cpp:43:3
    #8 0x31d9d281 in Poco::Net::TCPServerDispatcher::run() /build/obj-x86_64-linux-gnu/../contrib/poco/Net/src/TCPServerDispatcher.cpp:114:20
    #9 0x3206b5d5 in Poco::PooledThread::run() /build/obj-x86_64-linux-gnu/../contrib/poco/Foundation/src/ThreadPool.cpp:199:14
    #10 0x320657ad in Poco::ThreadImpl::runnableEntry(void*) /build/obj-x86_64-linux-gnu/../contrib/poco/Foundation/src/Thread_POSIX.cpp:345:27
    #11 0x7ffff7f853e8 in start_thread (/usr/lib/libpthread.so.0+0x93e8)
    #12 0x7ffff7ea2292 in clone (/usr/lib/libc.so.6+0x100292)

0x6120000223c0 is located 0 bytes inside of 272-byte region [0x6120000223c0,0x6120000224d0)
freed by thread T22 (TCPHandler) here:
    #0 0x122f3b62 in operator delete(void*, unsigned long) (/src/ch/tmp/master-20200831/clickhouse+0x122f3b62)
    #1 0x2bd9e9fa in std::__1::default_delete<DB::IQueryPlanStep>::operator()(DB::IQueryPlanStep*) const /build/obj-x86_64-linux-gnu/../contrib/libcxx/include/memory:2363:5
    #2 0x2bd9e9fa in std::__1::unique_ptr<DB::IQueryPlanStep, std::__1::default_delete<DB::IQueryPlanStep> >::reset(DB::IQueryPlanStep*) /build/obj-x86_64-linux-gnu/../contrib/libcxx/include/memory:2618:7
    #3 0x2bd9e9fa in std::__1::unique_ptr<DB::IQueryPlanStep, std::__1::default_delete<DB::IQueryPlanStep> >::~unique_ptr() /build/obj-x86_64-linux-gnu/../contrib/libcxx/include/memory:2572:19
    #4 0x2bd9e9fa in DB::QueryPlan::Node::~Node() /build/obj-x86_64-linux-gnu/../src/Processors/QueryPlan/QueryPlan.h:66:12
    #5 0x2bd9e9fa in void std::__1::allocator_traits<std::__1::allocator<std::__1::__list_node<DB::QueryPlan::Node, void*> > >::__destroy<DB::QueryPlan::Node>(std::__1::integral_constant<bool, false>, std::__1::allocator<std::__1::__list_node<DB::QueryPlan::Node, void*> >&,
 DB::QueryPlan::Node*) /build/obj-x86_64-linux-gnu/../contrib/libcxx/include/memory:1798:23
    #6 0x2bd9e9fa in void std::__1::allocator_traits<std::__1::allocator<std::__1::__list_node<DB::QueryPlan::Node, void*> > >::destroy<DB::QueryPlan::Node>(std::__1::allocator<std::__1::__list_node<DB::QueryPlan::Node, void*> >&, DB::QueryPlan::Node*) /build/obj-x86_64-lin
ux-gnu/../contrib/libcxx/include/memory:1630:14
    #7 0x2bd9e9fa in std::__1::__list_imp<DB::QueryPlan::Node, std::__1::allocator<DB::QueryPlan::Node> >::clear() /build/obj-x86_64-linux-gnu/../contrib/libcxx/include/list:762:13
    #8 0x29fece08 in DB::InterpreterSelectQuery::execute() /build/obj-x86_64-linux-gnu/../src/Interpreters/InterpreterSelectQuery.cpp:492:1
    #9 0x2abf7484 in DB::ClusterProxy::(anonymous namespace)::createLocalStream(std::__1::shared_ptr<DB::IAST> const&, DB::Block const&, DB::Context const&, DB::QueryProcessingStage::Enum) /build/obj-x86_64-linux-gnu/../src/Interpreters/ClusterProxy/SelectStreamFactory.cpp:
78:33
    #10 0x2abea85d in DB::ClusterProxy::SelectStreamFactory::createForShard(DB::Cluster::ShardInfo const&, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, std::__1::shared_ptr<DB::IAST> const&, DB::Context const&, std::__1::shar
ed_ptr<DB::Throttler> const&, DB::SelectQueryInfo const&, std::__1::vector<DB::Pipe, std::__1::allocator<DB::Pipe> >&)::$_0::operator()() const /build/obj-x86_64-linux-gnu/../src/Interpreters/ClusterProxy/SelectStreamFactory.cpp:133:51
    #11 0x2abea85d in DB::ClusterProxy::SelectStreamFactory::createForShard(DB::Cluster::ShardInfo const&, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, std::__1::shared_ptr<DB::IAST> const&, DB::Context const&, std::__1::shar
ed_ptr<DB::Throttler> const&, DB::SelectQueryInfo const&, std::__1::vector<DB::Pipe, std::__1::allocator<DB::Pipe> >&) /build/obj-x86_64-linux-gnu/../src/Interpreters/ClusterProxy/SelectStreamFactory.cpp:189:13
    #12 0x2abe6d99 in DB::ClusterProxy::executeQuery(DB::ClusterProxy::IStreamFactory&, std::__1::shared_ptr<DB::Cluster> const&, Poco::Logger*, std::__1::shared_ptr<DB::IAST> const&, DB::Context const&, DB::Settings const&, DB::SelectQueryInfo const&) /build/obj-x86_64-lin
ux-gnu/../src/Interpreters/ClusterProxy/executeQuery.cpp:107:24
    #13 0x2abc4b74 in DB::StorageDistributed::read(std::__1::vector<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::__1::allocator<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > > > const&, s
td::__1::shared_ptr<DB::StorageInMemoryMetadata const> const&, DB::SelectQueryInfo const&, DB::Context const&, DB::QueryProcessingStage::Enum, unsigned long, unsigned int) /build/obj-x86_64-linux-gnu/../src/Storages/StorageDistributed.cpp:514:12
    #14 0x2bda1c5a in DB::ReadFromStorageStep::ReadFromStorageStep(std::__1::shared_ptr<DB::RWLockImpl::LockHolderImpl>, std::__1::shared_ptr<DB::StorageInMemoryMetadata const>&, DB::SelectQueryOptions, std::__1::shared_ptr<DB::IStorage>, std::__1::vector<std::__1::basic_st
ring<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::__1::allocator<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > > > const&, DB::SelectQueryInfo const&, std::__1::shared_ptr<DB::Context>, DB::QueryProcessingStage
::Enum, unsigned long, unsigned long) /build/obj-x86_64-linux-gnu/../src/Processors/QueryPlan/ReadFromStorageStep.cpp:39:26
    #15 0x2a01ca70 in std::__1::__unique_if<DB::ReadFromStorageStep>::__unique_single std::__1::make_unique<DB::ReadFromStorageStep, std::__1::shared_ptr<DB::RWLockImpl::LockHolderImpl>&, std::__1::shared_ptr<DB::StorageInMemoryMetadata const>&, DB::SelectQueryOptions&, std
::__1::shared_ptr<DB::IStorage>&, std::__1::vector<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::__1::allocator<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > > >&, DB::SelectQueryInfo&, st
d::__1::shared_ptr<DB::Context>&, DB::QueryProcessingStage::Enum&, unsigned long&, unsigned long&>(std::__1::shared_ptr<DB::RWLockImpl::LockHolderImpl>&, std::__1::shared_ptr<DB::StorageInMemoryMetadata const>&, DB::SelectQueryOptions&, std::__1::shared_ptr<DB::IStorage>&,
std::__1::vector<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::__1::allocator<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > > >&, DB::SelectQueryInfo&, std::__1::shared_ptr<DB::Context>&,
DB::QueryProcessingStage::Enum&, unsigned long&, unsigned long&) /build/obj-x86_64-linux-gnu/../contrib/libcxx/include/memory:3028:32
    #16 0x29ff556a in DB::InterpreterSelectQuery::executeFetchColumns(DB::QueryProcessingStage::Enum, DB::QueryPlan&, std::__1::shared_ptr<DB::PrewhereInfo> const&, std::__1::vector<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, std::
__1::allocator<std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > > > const&) /build/obj-x86_64-linux-gnu/../src/Interpreters/InterpreterSelectQuery.cpp:1383:26
    #17 0x29fe6b83 in DB::InterpreterSelectQuery::executeImpl(DB::QueryPlan&, std::__1::shared_ptr<DB::IBlockInputStream> const&, std::__1::optional<DB::Pipe>) /build/obj-x86_64-linux-gnu/../src/Interpreters/InterpreterSelectQuery.cpp:795:9
    #18 0x29fe5771 in DB::InterpreterSelectQuery::buildQueryPlan(DB::QueryPlan&) /build/obj-x86_64-linux-gnu/../src/Interpreters/InterpreterSelectQuery.cpp:473:5
    #19 0x2a47d370 in DB::InterpreterSelectWithUnionQuery::buildQueryPlan(DB::QueryPlan&) /build/obj-x86_64-linux-gnu/../src/Interpreters/InterpreterSelectWithUnionQuery.cpp:182:38
    #20 0x29ee5bff in DB::InterpreterExplainQuery::executeImpl() /build/obj-x86_64-linux-gnu/../src/Interpreters/InterpreterExplainQuery.cpp:265:21
    #21 0x29ee2e40 in DB::InterpreterExplainQuery::execute() /build/obj-x86_64-linux-gnu/../src/Interpreters/InterpreterExplainQuery.cpp:73:14
    #22 0x2a7b44a2 in DB::executeQueryImpl(char const*, char const*, DB::Context&, bool, DB::QueryProcessingStage::Enum, bool, DB::ReadBuffer*) /build/obj-x86_64-linux-gnu/../src/Interpreters/executeQuery.cpp:389:28
    #23 0x2a7b1cb3 in DB::executeQuery(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, DB::Context&, bool, DB::QueryProcessingStage::Enum, bool) /build/obj-x86_64-linux-gnu/../src/Interpreters/executeQuery.cpp:675:30
    #24 0x2b7993b2 in DB::TCPHandler::runImpl() /build/obj-x86_64-linux-gnu/../src/Server/TCPHandler.cpp:253:24
    #25 0x2b7b649a in DB::TCPHandler::run() /build/obj-x86_64-linux-gnu/../src/Server/TCPHandler.cpp:1217:9
    #26 0x31d9c57e in Poco::Net::TCPServerConnection::start() /build/obj-x86_64-linux-gnu/../contrib/poco/Net/src/TCPServerConnection.cpp:43:3
    #27 0x31d9d281 in Poco::Net::TCPServerDispatcher::run() /build/obj-x86_64-linux-gnu/../contrib/poco/Net/src/TCPServerDispatcher.cpp:114:20
    #28 0x3206b5d5 in Poco::PooledThread::run() /build/obj-x86_64-linux-gnu/../contrib/poco/Foundation/src/ThreadPool.cpp:199:14
    #29 0x320657ad in Poco::ThreadImpl::runnableEntry(void*) /build/obj-x86_64-linux-gnu/../contrib/poco/Foundation/src/Thread_POSIX.cpp:345:27
    #30 0x7ffff7f853e8 in start_thread (/usr/lib/libpthread.so.0+0x93e8)
```

</details>