Protocol Documentation

display-friendly identification

computer-friendly identification

Starts a document in a given console - to end, just close the stream, the server will hang up right away

Modifies the document that autocomplete can be requested on

Requests that a response be sent back with completion items

Request for help about the method signature at the cursor

Request for help about what the user is hovering over

Request to perform file diagnostics

Closes the document, indicating that it will not be referenced again

Deprecated.

Deprecated.

Deprecated.

Deprecated.

Deprecated. Maintained for backwards compatibility. Use the same field on AutoCompleteResponse instead

Deprecated. Maintained for backwards compatibility. Use the same field on AutoCompleteResponse instead

Returns the maximum amount of memory that the Java virtual machine will attempt to use. If there is no inherent limit then the value Long.MAX_VALUE will be returned. the maximum amount of memory that the virtual machine will attempt to use, measured in bytes

Returns the total amount of memory in the Java virtual machine. The value returned by this method may vary over time, depending on the host environment. Note that the amount of memory required to hold an object of any given type may be implementation-dependent. the total amount of memory currently available for current and future objects, measured in bytes.

Returns the amount of free memory in the Java Virtual Machine. Calling the gc method may result in increasing the value returned by freeMemory. an approximation to the total amount of memory currently available for future allocated objects, measured in bytes.

Ticket console_id = 1; If a non-zero value is specified, represents the timestamp in microseconds since the unix epoch when the client last saw a message. Technically this might skip messages if more than one message was logged at the same microsecond that connection was lost - to avoid this, subtract one from the last seen message's micros, and expect to receive some messages that have already been seen.

Deprecated.

Ticket to use to hold the result HierarchicalTable (RollupTable or TreeTable) from the applying the operations

Ticket for the input HierarchicalTable (RollupTable or TreeTable) to apply operations to

Filters to apply to the input HierarchicalTable to produce the result HierarchicalTable. Never expressed against the "structural" columns included in the a HierarchicalTableDescriptor's snapshot_schema. For RollupTables, only the group-by columns may be filtered. The names are always expressed as they appear in aggregated node columns (and in the group-by columns). The filtering will result in a complete or partial new Table.rollup operation. For TreeTables, these may be variously applied to the source (resulting in a new Table.tree operation) or to the nodes (resulting in filtering at snapshot time).

Sorts to apply to the input HierarchicalTable to produce the result HierarchicalTable. Never expressed against the "structural" columns included in the a HierarchicalTableDescriptor's snapshot_schema. For TreeTables, these are simply applied to the nodes at snapshot time. For RollupTables, these are expressed against the aggregated node columns, and will be applied to the appropriate input (constituent) columns as well. The appropriate (aggregated or constituent) sorts are applied to the nodes at snapshot time.

Schema to be used for snapshot or subscription requests as described in Arrow Message.fbs::Message. Field metadata is used to convey additional information about the structure of the HierarchicalTable, the special roles some columns play, and the relationships between columns. "hierarchicalTable.isStructuralColumn" is always "true" if set, and is set on columns that should be included on every snapshot or subscription request, but should not be directly user-visible. "hierarchicalTable.isExpandByColumn" is always "true" if set, and is set on all the columns that must be included in a HierarchicalTableViewRequest's key table, if a key table is specified. These columns are generally user-visible and displayed before other columns, unless they also have "hierarchicalTable.isStructuralColumn" set. "hierarchicalTable.isRowDepthColumn" is always "true" if set, and is set on a single column that specifies the depth of a row. That column will always have "hierarchicalTable.isExpandByColumn" set for RollupTables, but never for TreeTables. "hierarchicalTable.isRowExpandedColumn" is always "true" if set, and is set on a single nullable column of booleans that specifies whether a row is expandable or expanded. Values will be null for rows that are not expandable, true for expanded rows, false for rows that are not expanded (but expandable). Leaf rows have no children to expand, and hence will always have a null value for this column. "rollupTable.isAggregatedNodeColumn" is always "true" if set, and is set on all columns of a RollupTable that belong to the aggregated nodes. "rollupTable.isConstituentNodeColumn" is always "true" if set, and is set on all columns of a RollupTable that belong to the constituent nodes. No such columns will be present if constituents are not included in the RollupTable. "rollupTable.isGroupByColumn" is always "true" if set, and is set on all columns of a RollupTable that are "group-by columns", whether the node is aggregated or constituent. All nodes have the same names and types for columns labeled in this way. Such columns will always have "hierarchicalTable.isExpandByColumn" set if and only if they also have "rollupTable.isAggregatedNodeColumn" set. "rollupTable.aggregationInputColumnName" is set to the (string) name of the corresponding constituent column that was used as input to this aggregation node column. May have an empty value, because some aggregations take no input columns, for example "Count". This is only ever present on columns with "rollupTable.isAggregatedNodeColumn" set. "treeTable.isNodeColumn" is always "true" if set, and is set on all columns of a TreeTable that nodes inherit from the source Table. "treeTable.isIdentifierColumn" is always "true" if set, and is set on the single column that uniquely identifies a TreeTable row and links it to its children. Such columns will always have "hierarchicalTable.isExpandByColumn" set. "treeTable.isParentIdentifierColumn" is always "true" if set, and is set on the single column that links a TreeTable row to its parent row.

Whether or not this table might change.

Ticket to use to hold an export of the HierarchicalTable's source Table

Ticket for the (existing) HierarchicalTable (RollupTable or TreeTable) to export the source Table for

Ticket that represents a Table of expanded or contracted keys from a HierarchicalTable (RollupTable or TreeTable). The format for the key Table is dictated by the schema from the corresponding HierarchicalTableDescriptor. It is expected to have one column for each "expand-by column", including the "row depth column" for RollupTables only, and (optionally) an "action" column whose name is specified in the key_table_action_column field. If the Table is empty the result will have only default nodes expanded.

The name of a column of bytes found in the key table that specifies the action desired for the node selected by the other columns for each row. Takes on the value 1 for nodes that should be expanded, 3 for nodes that should be expanded along with their descendants, and 4 for nodes that should be contracted. If this column name is not present, all nodes in the key table will be expanded without their descendants.

Ticket to use to hold the result HierarchicalTableView

Ticket for the HierarchicalTable (RollupTable or TreeTable) to expand

Ticket for an existing HierarchicalTableView. The result view will inherit the HierarchicalTable from the existing view. The two views will share state used for caching snapshot data, but the server implementation may limit parallelism when performing snapshots for either view. Use this field when you intend to stop using the existing view and instead begin to use the result view.

Description for the expansions that define this view of the HierarchicalTable. If not present, the result will have default expansions, For RollupTables this will be the root (single row, top-level aggregation) and the next level if one exists (that is, if there are one or more group-by columns, or constituents are included). For TreeTables, this will be the root (one row for each child of the "null" parent identifier).

Ticket to use to hold the result RollupTable from the rollup operation

Ticket for the source Table to rollup

The aggregations that should be applied at each level of the rollup

Whether to include the leaf-level constituents in the result

The names of the columns to rollup by

Ticket to use to hold the result TreeTable from the tree operation

Ticket for the source Table to tree

The name of the column containing the unique identifier for each row in the source table

The name of the column containing the parent row's unique identifier for each row in the source table

Whether to promote "orphaned" nodes to be children of the root node. Orphans are nodes whose parent identifiers do not occur as identifiers for any row in the source Table.

The payload, may be empty.

The typed references, may be empty. These references may be any ticket, resolved or not. This lets the client reference objects that already exist on the server or are still pending. Note that pending tickets require the server to wait until that object exists before passing this request to the server plugin, and since messages are always processed in order, later requests will also be delayed.

The payload, may be empty.

The exported references, may be empty. To correctly free up unused server resources, clients must take care to release these exports when they will no longer be used. A reference may be missing a type, meaning that the object cannot be used as the source_id for a ConnectRequest, but it may still be passed back to the server as part of ClientData references, and it still needs to be released when no longer used.

Indicates that this is the first request of the stream, asking to connect to a specific object on the server.

Data to pass to the object on the server.

Data to pass to the client about the object on the server.

The ticket for the PartitionedTable object to query.

The ticket for the table containing the key to fetch from the partitioned table.

The ticket to use to hold the newly returned table.

The ticket for the PartitionedTable object to merge.

The ticket to use to hold the results of the merge operation.

The names of the key columns. The underlying table will contain these columns - a client can subscribe to these columns to see what keys are present.

The name of the column in the underlying table that contains the table represented by that row.

True if the keys will be unique, so any set of known keys can be queried using GetTable.

Returns a flight Messsage wrapping a Schema that will describe every table contained in this PartitionedTable.

True if the underlying table may tick with updates. See PartitionedTable.constituentChangesPermitted() for more details.

any errors will include an id that can be used to find details of the error in the logs

will be set to an identifier of the dependency that cascaded the error if applicable

Deprecated. A defined protocol version. Deephaven's OSS protocols are as follows: - protocol = 0: most recent HandshakeResponse payload - protocol = 1: payload is BasicAuth

Deprecated. Arbitrary auth/handshake info.

Deprecated. The metadata header to identify the session. This value is static and defined via configuration.

Deprecated. Arbitrary session_token to assign to the value to the provided metadata header.

Deprecated. When this session_token will be considered invalid by the server.

Deprecated. The length of time that this token was intended to live. Note that `refreshSessionToken` may return the existing token to reduce overhead and to prevent denial-of-service caused by refreshing too frequently.

whether or not this termination is expected

if additional information is available then provide it in this field

if this is due to an exception, whether or not it was uncaught

if applicable, the list of stack traces in reverse causal order

The type of the protobuf the auth payload protobuf.

The serialized payload of the protobuf instance.

The path to the directory to create

The path to the item to delete.

The path to the file to read

If present, tells the server to not send a result if the etag matches the current file's content.

Contains the contents of the file, unless the returned etag matches the requested etag.

Represents the current etag of the requested file. If an etag was in the request and this matches, contents should be ignored, and the existing client copy of the file is already correct. In all other cases, this etag can be used in future requests to see if the file's contents are different.

The path to the item that this message describes.

The type of this item, either file or directory.

If this message represents a file, this is the size of the file.

Opaque string value representing a hash of the contents of this file, if available.

The path to the directory to list. empty to list top level

A pattern to filter for, with "?" to match any one character, "*" to match any number of characters, and "{}"s to hold a comma-separated list of possible matches. The format follows Java's FileSystem.getPathMatcher (see https://docs.oracle.com/javase/8/docs/api/java/nio/file/FileSystem.html#getPathMatcher-java.lang.String-), except without allowing subdirectories with / or **.

List of items found in the specified directory.

The canonical path of the listed directory. This is useful to recognize the basename of the items in a cross-platform way.

The path where the file currently exists

The path where the file should be moved to

True to permit replacing an existing file, false to require that no file already exists with that name.

True to permit replacing an existing file, false to require that no file already exists with that name.

The path to the file to write contents to

The contents to use when creating then file, or to use to replace the file.

Represents the etag of the saved contents, so the client can check for external changes.

Percentile. Must be in range [0.0, 1.0].

T-Digest compression factor. Must be greater than or equal to 1. 1000 is extremely large. When not specified, the server will choose a compression value.

Whether null input values should be included when counting the distinct input values.

Whether null input values should be included in the distinct output values.

The formula to use to calculate output values from grouped input values.

The formula parameter token to be replaced with the input column name for evaluation.

Whether to average the highest low-bucket value and lowest high-bucket value, when the low-bucket and high-bucket are of equal size. Only applies to numeric types.

Expected to be in range [Byte.MIN_VALUE, Byte.MAX_VALUE]

Expected to be in range [Short.MIN_VALUE, Short.MAX_VALUE]

Expected to be in range [0x0000, 0xFFFF] TODO(deephaven-core#3212): Expand AggSpecNonUniqueSentinel types

The percentile to calculate. Must be in the range [0.0, 1.0].

Whether to average the highest low-bucket value and lowest high-bucket value, when the low-bucket and high-bucket are of equal size. Only applies to numeric types.

Using a message instead of string to support backwards-compatibility in the future

TODO(deephaven-core#821): SortedFirst / SortedLast aggregations with sort direction

T-Digest compression factor. Must be greater than or equal to 1. 1000 is extremely large. When not specified, the server will choose a compression value.

Whether to include null values as a distinct value for determining if there is only one unique value to output

The output value to use for groups that don't have a single unique input value

Column name for the source of input weights.

A table whose distinct combinations of values for the group_by_columns should be used to create an initial set of aggregation groups. All other columns are ignored. This is useful in combination with preserve_empty == true to ensure that particular groups appear in the result table, or with preserve_empty == false to control the encounter order for a collection of groups and thus their relative order in the result. Changes to initial_group_ids are not expected or handled; if initial_groups_id is a refreshing table, only its contents at instantiation time will be used. If initial_groups_id is not present, the result will be the same as if a table with no rows was supplied.

Whether to keep result rows for groups that are initially empty or become empty as a result of updates. Each aggregation operator defines its own value for empty groups.

The output column name

This is a comparison expression for the inexact as-of join match. In the case of an as-of join (aj), the comparison operator can be either ">=" or ">"; for example, "Foo>=Bar" or "Foo>Bar". In the case of a reverse-as-of join (raj), the comparison operator can be either "<=" or "<"; for example, "Foo<=Bar" or "Foo<Bar". In the case where the column name exists in both tables, the single column name can be used and it will inherit the default comparison operator: in the aj case, "Foo" is equivalent to "Foo>=Foo"; in the raj case, "Foo" is equivalent to "Foo<=Foo".

Direction to search to find a match. LESS_THAN_EQUAL and LESS_THAN will be used to make a Table.aj() call, and GREATER_THAN_EQUAL and GREATER_THAN will be used to make a Table.raj() call.

Deprecated.

The name of the column in the source table to read when generating statistics.

For non-numeric, non-date types, specify the max number of unique values to return, sorted by popularity. Leave unset to use server default, specify zero to skip.

don't use direct single-aggregate table operations even if there is only a single aggregate

used in all aggregates except countBy

countBy result (output) column OR weighted avg weight (input) column, otherwise unused

required by percentileBy aggregates, otherwise unused

used in percentileBy only

Optional, either this or schema must be specified, not both.

Schema as described in Arrow Message.fbs::Message. Optional, either this or source_table_id must be specified.

Specifies what type of input table to create.

the number of bits of key-space to initially reserve per group; default is 10

If this is part of a batch, you may receive creation messages that indicate the sub-operation failed.

If this is part of a batch, this errorInfo will be the message provided

Schema as described in Arrow Message.fbs::Message.

Whether or not this table might change.

The current number of rows for this table. If this is negative, the table isn't coalesced, meaning the size isn't known without scanning partitions. Typically, the client should filter the data by the partitioning columns first.

nanos since the epoch

if specified, the result will be sorted by this column

The source table to include in the multi-join output table.

The key columns to match; may be renamed to match other source table key columns.

The columns from the source table to include; if not provided, all columns are included.

The source table input specifications. One or more must be provided.

Provide detailed range match parameters for the range join (alternative to providing `range_match`)

Specifies the range match parameters as a parseable string. Providing `range_match` in the GRPC call is the alternative to detailed range match parameters provided in the `left_start_column`, `range_start_rule`, `right_range_column`, `range_end_rule`, and `left_end_column` fields.