Storage provider that reads ntuple pages from a file.
Definition at line 124 of file RPageStorageFile.hxx.
Classes | |
| struct | RFileCounters |
| File-specific I/O performance counters. More... | |
| struct | RStructureBuffer |
| Holds the uncompressed header and footer. More... | |
Public Types | |
| using | ColumnHandle_t = RColumnHandle |
| The column handle identifies a column with the current open page storage. | |
| using | SealedPageSequence_t = std::deque<RSealedPage> |
Public Member Functions | |
| RPageSourceFile (const RPageSourceFile &)=delete | |
| RPageSourceFile (RPageSourceFile &&)=delete | |
| RPageSourceFile (std::string_view ntupleName, std::string_view path, const ROOT::RNTupleReadOptions &options) | |
| RPageSourceFile (std::string_view ntupleName, std::unique_ptr< RRawFile > file, const ROOT::RNTupleReadOptions &options) | |
| ~RPageSourceFile () override | |
| ColumnHandle_t | AddColumn (ROOT::DescriptorId_t fieldId, ROOT::Internal::RColumn &column) override |
| Register a new column. | |
| void | Attach (ROOT::Internal::RNTupleSerializer::EDescriptorDeserializeMode mode=ROOT::Internal::RNTupleSerializer::EDescriptorDeserializeMode::kForReading) |
| Open the physical storage container and deserialize header and footer. | |
| std::unique_ptr< RPageSource > | Clone () const |
| Open the same storage multiple time, e.g. | |
| void | DropColumn (ColumnHandle_t columnHandle) override |
| Unregisters a column. | |
| ROOT::DescriptorId_t | GetColumnId (ColumnHandle_t columnHandle) const |
| REntryRange | GetEntryRange () const |
| virtual ROOT::Experimental::Detail::RNTupleMetrics & | GetMetrics () |
| Returns the default metrics object. | |
| ROOT::NTupleSize_t | GetNElements (ColumnHandle_t columnHandle) |
| ROOT::NTupleSize_t | GetNEntries () |
| const std::string & | GetNTupleName () const |
| Returns the NTuple name. | |
| const std::unordered_set< ROOT::DescriptorId_t > & | GetPinnedClusters () const |
| const ROOT::RNTupleReadOptions & | GetReadOptions () const |
| const RSharedDescriptorGuard | GetSharedDescriptorGuard () const |
| Takes the read lock for the descriptor. | |
| EPageStorageType | GetType () final |
| Whether the concrete implementation is a sink or a source. | |
| std::vector< std::unique_ptr< ROOT::Internal::RCluster > > | LoadClusters (std::span< ROOT::Internal::RCluster::RKey > clusterKeys) final |
| virtual ROOT::Internal::RPageRef | LoadPage (ColumnHandle_t columnHandle, RNTupleLocalIndex localIndex) |
| Another version of LoadPage that allows to specify cluster-relative indexes. | |
| virtual ROOT::Internal::RPageRef | LoadPage (ColumnHandle_t columnHandle, ROOT::NTupleSize_t globalIndex) |
| Allocates and fills a page that contains the index-th element. | |
| void | LoadSealedPage (ROOT::DescriptorId_t physicalColumnId, RNTupleLocalIndex localIndex, RSealedPage &sealedPage) final |
| Read the packed and compressed bytes of a page into the memory buffer provided by sealedPage. | |
| void | LoadStreamerInfo () final |
| Forces the loading of ROOT StreamerInfo from the underlying file. | |
| void | LoadStructure () |
| Loads header and footer without decompressing or deserializing them. | |
| std::unique_ptr< RPageSource > | OpenWithDifferentAnchor (const ROOT::Internal::RNTupleLink &anchorLink, const ROOT::RNTupleReadOptions &options={}) final |
| Creates a new PageSource using the same underlying file as this but referring to a different RNTuple, described by anchorLink. | |
| RPageSourceFile & | operator= (const RPageSourceFile &)=delete |
| RPageSourceFile & | operator= (RPageSourceFile &&)=delete |
| void | PinCluster (ROOT::DescriptorId_t clusterId) |
| Instructs the cluster pool and page pool to consider the given cluster as active (should stay cached). | |
| void | RegisterStreamerInfos () |
| Builds the streamer info records from the descriptor's extra type info section. | |
| void | SetEntryRange (const REntryRange &range) |
| Promise to only read from the given entry range. | |
| void | SetTaskScheduler (RTaskScheduler *taskScheduler) |
| void | UnpinCluster (ROOT::DescriptorId_t clusterId) |
| Allows the given cluster to be evicted from the cluster pool and page pool. | |
| RResult< ROOT::Internal::RPage > | UnsealPage (const RSealedPage &sealedPage, const ROOT::Internal::RColumnElementBase &element) |
| void | UnzipCluster (ROOT::Internal::RCluster *cluster) |
| Populates all the pages of the given cluster ids and columns; it is possible that some columns do not contain any pages. | |
Static Public Member Functions | |
| static std::unique_ptr< RPageSource > | Create (std::string_view ntupleName, std::string_view location, const ROOT::RNTupleReadOptions &options=ROOT::RNTupleReadOptions()) |
| Guess the concrete derived page source from the file name (location). | |
| static std::unique_ptr< RPageSourceFile > | CreateFromAnchor (const RNTuple &anchor, const ROOT::RNTupleReadOptions &options=ROOT::RNTupleReadOptions()) |
| Used from the RNTuple class to build a datasource if the anchor is already available. | |
| static RResult< ROOT::Internal::RPage > | UnsealPage (const RSealedPage &sealedPage, const ROOT::Internal::RColumnElementBase &element, ROOT::Internal::RPageAllocator &pageAlloc) |
| Helper for unstreaming a page. | |
Static Public Attributes | |
| static constexpr std::size_t | kNBytesPageChecksum = sizeof(std::uint64_t) |
| The page checksum is a 64bit xxhash3. | |
Protected Member Functions | |
| ROOT::RNTupleDescriptor | AttachImpl (RNTupleSerializer::EDescriptorDeserializeMode mode) final |
| LoadStructureImpl() has been called before AttachImpl() is called | |
| std::unique_ptr< RPageSource > | CloneImpl () const final |
| The cloned page source creates a new raw file and reader and opens its own file descriptor to the data. | |
| void | EnableDefaultMetrics (const std::string &prefix) |
| Enables the default set of metrics provided by RPageSource. | |
| RExclDescriptorGuard | GetExclDescriptorGuard () |
| Note that the underlying lock is not recursive. See GetSharedDescriptorGuard() for further information. | |
| RPageRef | LoadPageImpl (ColumnHandle_t columnHandle, const RClusterInfo &clusterInfo, ROOT::NTupleSize_t idxInCluster) final |
| void | LoadStructureImpl () final |
| void | PrepareLoadCluster (const ROOT::Internal::RCluster::RKey &clusterKey, ROOT::Internal::ROnDiskPageMap &pageZeroMap, std::function< void(ROOT::DescriptorId_t, ROOT::NTupleSize_t, const ROOT::RClusterDescriptor::RPageInfo &)> perPageFunc) |
| Prepare a page range read for the column set in clusterKey. | |
| virtual void | UnzipClusterImpl (ROOT::Internal::RCluster *cluster) |
| void | WaitForAllTasks () |
Protected Attributes | |
| RActivePhysicalColumns | fActivePhysicalColumns |
| The active columns are implicitly defined by the model fields or views. | |
| ROOT::Internal::RClusterPool | fClusterPool |
| The cluster pool asynchronously preloads the next few clusters. | |
| std::unique_ptr< RCounters > | fCounters |
| ROOT::Experimental::Detail::RNTupleMetrics | fMetrics |
| std::string | fNTupleName |
| ROOT::RNTupleReadOptions | fOptions |
| std::unique_ptr< ROOT::Internal::RPageAllocator > | fPageAllocator |
| For the time being, we will use the heap allocator for all sources and sinks. This may change in the future. | |
| ROOT::Internal::RPagePool | fPagePool |
| Pages that are unzipped with IMT are staged into the page pool. | |
| std::unordered_set< ROOT::DescriptorId_t > | fPinnedClusters |
| Pinned clusters and their $2 * (cluster bunch size) - 1$ successors will not be evicted from the cluster pool. | |
| RTaskScheduler * | fTaskScheduler = nullptr |
Private Member Functions | |
| RPageSourceFile (std::string_view ntupleName, const ROOT::RNTupleReadOptions &options) | |
| std::unique_ptr< ROOT::Internal::RCluster > | PrepareSingleCluster (const ROOT::Internal::RCluster::RKey &clusterKey, std::vector< RRawFile::RIOVec > &readRequests) |
| Helper function for LoadClusters: it prepares the memory buffer (page map) and the read requests for a given cluster and columns. | |
| void | UpdateLastUsedCluster (ROOT::DescriptorId_t clusterId) |
| Does nothing if fLastUsedCluster == clusterId. | |
Private Attributes | |
| std::optional< RNTuple > | fAnchor |
| Either provided by CreateFromAnchor, or read from the ROOT file given the ntuple name. | |
| ROOT::Internal::RCluster * | fCurrentCluster = nullptr |
| The last cluster from which a page got loaded. Points into fClusterPool->fPool. | |
| ROOT::RNTupleDescriptor | fDescriptor |
| RNTupleDescriptorBuilder | fDescriptorBuilder |
| The descriptor is created from the header and footer either in AttachImpl or in CreateFromAnchor. | |
| std::shared_mutex | fDescriptorLock |
| REntryRange | fEntryRange |
| Used by the cluster pool to prevent reading beyond the given range. | |
| std::unique_ptr< RRawFile > | fFile |
| An RRawFile is used to request the necessary byte ranges from a local or a remote file. | |
| std::unique_ptr< RFileCounters > | fFileCounters |
| std::int64_t | fFileSize = 0 |
| Total file size, set once in AttachImpl(). | |
| bool | fHasStreamerInfosRegistered = false |
| Set to true when RegisterStreamerInfos() is called. | |
| bool | fHasStructure = false |
| Set to true once LoadStructure() is called. | |
| bool | fIsAttached = false |
| Set to true once Attach() is called. | |
| std::uint64_t | fLastOffset = 0 |
| Tracks the last read offset for seek distance calculation. | |
| ROOT::DescriptorId_t | fLastUsedCluster = ROOT::kInvalidDescriptorId |
| Remembers the last cluster id from which a page was requested. | |
| std::map< ROOT::NTupleSize_t, ROOT::DescriptorId_t > | fPreloadedClusters |
| Clusters from where pages got preloaded in UnzipClusterImpl(), ordered by first entry number of the clusters. | |
| ROOT::Internal::RMiniFileReader | fReader |
| Takes the fFile to read ntuple blobs from it. | |
| RStructureBuffer | fStructureBuffer |
| Populated by LoadStructureImpl(), reset at the end of Attach(). | |
Friends | |
| class | ROOT::RNTuple |
#include <ROOT/RPageStorageFile.hxx>
|
inherited |
The column handle identifies a column with the current open page storage.
Definition at line 180 of file RPageStorage.hxx.
|
inherited |
Definition at line 130 of file RPageStorage.hxx.
|
private |
Definition at line 326 of file RPageStorageFile.cxx.
| ROOT::Internal::RPageSourceFile::RPageSourceFile | ( | std::string_view | ntupleName, |
| std::string_view | path, | ||
| const ROOT::RNTupleReadOptions & | options ) |
Definition at line 382 of file RPageStorageFile.cxx.
| ROOT::Internal::RPageSourceFile::RPageSourceFile | ( | std::string_view | ntupleName, |
| std::unique_ptr< RRawFile > | file, | ||
| const ROOT::RNTupleReadOptions & | options ) |
Definition at line 372 of file RPageStorageFile.cxx.
|
delete |
|
override |
Definition at line 415 of file RPageStorageFile.cxx.
|
overridevirtualinherited |
Register a new column.
When reading, the column must exist in the ntuple on disk corresponding to the metadata. When writing, every column can only be attached once.
Implements ROOT::Internal::RPageStorage.
Definition at line 197 of file RPageStorage.cxx.
|
inherited |
Open the physical storage container and deserialize header and footer.
Definition at line 227 of file RPageStorage.cxx.
|
finalprotectedvirtual |
LoadStructureImpl() has been called before AttachImpl() is called
Implements ROOT::Internal::RPageSource.
Definition at line 480 of file RPageStorageFile.cxx.
|
inherited |
Open the same storage multiple time, e.g.
for reading in multiple threads. If the source is already attached, the clone will be attached, too. The clone will use, however, it's own connection to the underlying storage (e.g., file descriptor, XRootD handle, etc.)
Definition at line 235 of file RPageStorage.cxx.
|
finalprotectedvirtual |
The cloned page source creates a new raw file and reader and opens its own file descriptor to the data.
Implements ROOT::Internal::RPageSource.
Definition at line 625 of file RPageStorageFile.cxx.
|
staticinherited |
Guess the concrete derived page source from the file name (location).
Definition at line 177 of file RPageStorage.cxx.
|
static |
Used from the RNTuple class to build a datasource if the anchor is already available.
Requires the RNTuple object to be streamed from a file.
Definition at line 389 of file RPageStorageFile.cxx.
|
overridevirtualinherited |
Unregisters a column.
A page source decreases the reference counter for the corresponding active column. For a page sink, dropping columns is currently a no-op.
Implements ROOT::Internal::RPageStorage.
Definition at line 207 of file RPageStorage.cxx.
|
protectedinherited |
Enables the default set of metrics provided by RPageSource.
prefix will be used as the prefix for the counters registered in the internal RNTupleMetrics object. A subclass using the default set of metrics is responsible for updating the counters appropriately, e.g. fCounters->fNRead.Inc() Alternatively, a subclass might provide its own RNTupleMetrics object by overriding the GetMetrics() member function.
Definition at line 471 of file RPageStorage.cxx.
|
inlineinherited |
Definition at line 188 of file RPageStorage.hxx.
|
inlineinherited |
Definition at line 805 of file RPageStorage.hxx.
|
inlineprotectedinherited |
Note that the underlying lock is not recursive. See GetSharedDescriptorGuard() for further information.
Definition at line 750 of file RPageStorage.hxx.
|
inlinevirtualinherited |
Returns the default metrics object.
Subclasses might alternatively provide their own metrics object by overriding this.
Definition at line 192 of file RPageStorage.hxx.
|
inherited |
Definition at line 251 of file RPageStorage.cxx.
|
inherited |
Definition at line 246 of file RPageStorage.cxx.
|
inlineinherited |
Returns the NTuple name.
Definition at line 195 of file RPageStorage.hxx.
|
inlineinherited |
Definition at line 843 of file RPageStorage.hxx.
|
inlineinherited |
Definition at line 775 of file RPageStorage.hxx.
|
inlineinherited |
Takes the read lock for the descriptor.
Multiple threads can take the lock concurrently. The underlying std::shared_mutex, however, is neither read nor write recursive: within one thread, only one lock (shared or exclusive) must be acquired at the same time. This requires special care in sections protected by GetSharedDescriptorGuard() and GetExclDescriptorGuard() especially to avoid that the locks are acquired indirectly. As a general guideline, no other method of the page source should be called (directly or indirectly) in a guarded section.
Definition at line 783 of file RPageStorage.hxx.
|
inlinefinalvirtualinherited |
Whether the concrete implementation is a sink or a source.
Implements ROOT::Internal::RPageStorage.
Definition at line 774 of file RPageStorage.hxx.
|
final |
Definition at line 759 of file RPageStorageFile.cxx.
|
virtualinherited |
Another version of LoadPage that allows to specify cluster-relative indexes.
Returns a default-constructed RPage for suppressed columns.
Definition at line 435 of file RPageStorage.cxx.
|
virtualinherited |
Allocates and fills a page that contains the index-th element.
The default implementation searches the page and calls LoadPageImpl(). Returns a default-constructed RPage for suppressed columns.
Definition at line 396 of file RPageStorage.cxx.
|
finalprotectedvirtual |
Implements ROOT::Internal::RPageSource.
Definition at line 550 of file RPageStorageFile.cxx.
|
finalvirtual |
Read the packed and compressed bytes of a page into the memory buffer provided by sealedPage.
The sealed page can be used subsequently in a call to RPageSink::CommitSealedPage. The fSize and fNElements member of the sealedPage parameters are always set. If sealedPage.fBuffer is nullptr, no data will be copied but the returned size information can be used by the caller to allocate a large enough buffer and call LoadSealedPage again.
Implements ROOT::Internal::RPageSource.
Definition at line 521 of file RPageStorageFile.cxx.
|
finalvirtual |
Forces the loading of ROOT StreamerInfo from the underlying file.
This currently only has an effect for TFile-backed sources.
Implements ROOT::Internal::RPageSource.
Definition at line 826 of file RPageStorageFile.cxx.
|
inherited |
Loads header and footer without decompressing or deserializing them.
This can be used to asynchronously open a file in the background. The method is idempotent and it is called as a first step in Attach(). Pages sources may or may not make use of splitting loading and processing metadata. Therefore, LoadStructure() may do nothing and defer loading the metadata to Attach().
Definition at line 220 of file RPageStorage.cxx.
|
finalprotectedvirtual |
Implements ROOT::Internal::RPageSource.
Definition at line 435 of file RPageStorageFile.cxx.
|
finalvirtual |
Creates a new PageSource using the same underlying file as this but referring to a different RNTuple, described by anchorLink.
Implements ROOT::Internal::RPageSource.
Definition at line 421 of file RPageStorageFile.cxx.
|
delete |
|
delete |
|
inlineinherited |
Instructs the cluster pool and page pool to consider the given cluster as active (should stay cached).
Definition at line 840 of file RPageStorage.hxx.
|
protectedinherited |
Prepare a page range read for the column set in clusterKey.
Specifically, pages referencing the kTypePageZero locator are filled in pageZeroMap; otherwise, perPageFunc is called for each page. This is commonly used as part of LoadClusters() in derived classes.
Definition at line 334 of file RPageStorage.cxx.
|
private |
Helper function for LoadClusters: it prepares the memory buffer (page map) and the read requests for a given cluster and columns.
The reead requests are appended to the provided vector. This way, requests can be collected for multiple clusters before sending them to RRawFile::ReadV().
Definition at line 634 of file RPageStorageFile.cxx.
|
inherited |
Builds the streamer info records from the descriptor's extra type info section.
This is necessary when connecting streamer fields so that emulated classes can be read.
Definition at line 608 of file RPageStorage.cxx.
|
inherited |
Promise to only read from the given entry range.
If set, prevents the cluster pool from reading-ahead beyond the given range. The range needs to be within [0, GetNEntries()).
Definition at line 212 of file RPageStorage.cxx.
|
inlineinherited |
Definition at line 197 of file RPageStorage.hxx.
|
inlineinherited |
Allows the given cluster to be evicted from the cluster pool and page pool.
Definition at line 842 of file RPageStorage.hxx.
|
inherited |
Definition at line 564 of file RPageStorage.cxx.
|
staticinherited |
Helper for unstreaming a page.
This is commonly used in derived, concrete page sources. The implementation currently always makes a memory copy, even if the sealed page is uncompressed and in the final memory layout. The optimization of directly mapping pages is left to the concrete page source implementations.
Definition at line 569 of file RPageStorage.cxx.
|
inherited |
Populates all the pages of the given cluster ids and columns; it is possible that some columns do not contain any pages.
The page source may load more columns than the minimal necessary set from columns. To indicate which columns have been loaded, LoadClusters()`` must mark them with SetColumnAvailable(). That includes the ones from the columns that don't have pages; otherwise subsequent requests for the cluster would assume an incomplete cluster and trigger loading again. LoadClusters()` is typically called from the I/O thread of a cluster pool, i.e. the method runs concurrently to other methods of the page source. */ virtual std::vector<std::unique_ptr<ROOT::Internal::RCluster>> LoadClusters(std::span<ROOT::Internal::RCluster::RKey> clusterKeys) = 0;
/** Parallel decompression and unpacking of the pages in the given cluster. The unzipped pages are supposed to be preloaded in a page pool attached to the source. The method is triggered by the cluster pool's unzip thread. It is an optional optimization, the method can safely do nothing. In particular, the actual implementation will only run if a task scheduler is set. In practice, a task scheduler is set if implicit multi-threading is turned on.
Definition at line 256 of file RPageStorage.cxx.
|
protectedvirtualinherited |
Definition at line 262 of file RPageStorage.cxx.
|
privateinherited |
Does nothing if fLastUsedCluster == clusterId.
Otherwise, updated fLastUsedCluster and evict unused paged from the page pool of all previous clusters. Must not be called when the descriptor guard is taken.
Definition at line 361 of file RPageStorage.cxx.
|
inlineprotectedinherited |
Definition at line 153 of file RPageStorage.hxx.
|
friend |
Definition at line 125 of file RPageStorageFile.hxx.
|
protectedinherited |
The active columns are implicitly defined by the model fields or views.
Definition at line 710 of file RPageStorage.hxx.
|
private |
Either provided by CreateFromAnchor, or read from the ROOT file given the ntuple name.
Definition at line 143 of file RPageStorageFile.hxx.
|
protectedinherited |
The cluster pool asynchronously preloads the next few clusters.
Note that derived classes should call fClusterPool.StopBackgroundThread() in their destructor so that the I/O background thread does not call methods from the destructed derived class.
Definition at line 718 of file RPageStorage.hxx.
|
protectedinherited |
Definition at line 706 of file RPageStorage.hxx.
|
private |
The last cluster from which a page got loaded. Points into fClusterPool->fPool.
Definition at line 145 of file RPageStorageFile.hxx.
|
privateinherited |
Definition at line 624 of file RPageStorage.hxx.
The descriptor is created from the header and footer either in AttachImpl or in CreateFromAnchor.
Definition at line 151 of file RPageStorageFile.hxx.
|
mutableprivateinherited |
Definition at line 625 of file RPageStorage.hxx.
|
privateinherited |
Used by the cluster pool to prevent reading beyond the given range.
Definition at line 626 of file RPageStorage.hxx.
|
private |
An RRawFile is used to request the necessary byte ranges from a local or a remote file.
Definition at line 147 of file RPageStorageFile.hxx.
|
private |
Definition at line 164 of file RPageStorageFile.hxx.
|
private |
Total file size, set once in AttachImpl().
Definition at line 166 of file RPageStorageFile.hxx.
|
privateinherited |
Set to true when RegisterStreamerInfos() is called.
Definition at line 629 of file RPageStorage.hxx.
|
privateinherited |
Set to true once LoadStructure() is called.
Definition at line 627 of file RPageStorage.hxx.
|
privateinherited |
Set to true once Attach() is called.
Definition at line 628 of file RPageStorage.hxx.
|
private |
Tracks the last read offset for seek distance calculation.
Definition at line 155 of file RPageStorageFile.hxx.
|
privateinherited |
Remembers the last cluster id from which a page was requested.
Definition at line 632 of file RPageStorage.hxx.
|
protectedinherited |
Definition at line 146 of file RPageStorage.hxx.
|
protectedinherited |
Definition at line 151 of file RPageStorage.hxx.
|
protectedinherited |
Definition at line 708 of file RPageStorage.hxx.
|
protectedinherited |
For the time being, we will use the heap allocator for all sources and sinks. This may change in the future.
Definition at line 149 of file RPageStorage.hxx.
|
protectedinherited |
Pages that are unzipped with IMT are staged into the page pool.
Definition at line 720 of file RPageStorage.hxx.
|
protectedinherited |
Pinned clusters and their $2 * (cluster bunch size) - 1$ successors will not be evicted from the cluster pool.
Pages of pinned clusters won't be evicted from the page pool.
Definition at line 713 of file RPageStorage.hxx.
|
privateinherited |
Clusters from where pages got preloaded in UnzipClusterImpl(), ordered by first entry number of the clusters.
If the last used cluster changes in LoadPage(), all unused pages from previous clusters are evicted from the page pool. Pinned clusters won't be evicted.
Definition at line 636 of file RPageStorage.hxx.
Takes the fFile to read ntuple blobs from it.
Definition at line 149 of file RPageStorageFile.hxx.
Populated by LoadStructureImpl(), reset at the end of Attach().
Definition at line 153 of file RPageStorageFile.hxx.
|
protectedinherited |
Definition at line 152 of file RPageStorage.hxx.
|
staticconstexprinherited |
The page checksum is a 64bit xxhash3.
Definition at line 73 of file RPageStorage.hxx.