Package 'AzureKusto'

Title: Interface to 'Kusto'/'Azure Data Explorer'
Description: An interface to 'Azure Data Explorer', also known as 'Kusto', a fast, distributed data exploration service from Microsoft: <https://azure.microsoft.com/en-us/products/data-explorer/>. Includes 'DBI' and 'dplyr' interfaces, with the latter modelled after the 'dbplyr' package, whereby queries are translated from R into the native 'KQL' query language and executed lazily. On the admin side, the package extends the object framework provided by 'AzureRMR' to support creation and deletion of databases, and management of database principals. Part of the 'AzureR' family of packages.
Authors: Hong Ooi [aut], Alex Kyllo [aut, cre], dbplyr development team [cph] (Original framework for dplyr/database interface), Microsoft [cph]
Maintainer: Alex Kyllo <[email protected]>
License: MIT + file LICENSE
Version: 1.1.3.9000
Built: 2024-12-24 03:58:34 UTC
Source: https://github.com/azure/azurekusto

Help Index

Append a join operation to the tbl_kusto object's ops list

Description

Append a join operation to the tbl_kusto object's ops list

Usage

add_op_join(
  type,
  x,
  y,
  by = NULL,
  suffix = NULL,
  .strategy = NULL,
  .shufflekeys = NULL,
  .num_partitions = NULL,
  .remote = NULL
)

Arguments

type

The name of the join type, one of: inner_join, left_join, right_join, full_join, semi_join, anti_join
x

The "left" tbl
y

The "right" tbl
by

A vector of column names; keys by which tbl x and tbl y will be joined
suffix

A vector of strings that will be appended to the names of non-join key columns that exist in both tbl x and tbl y to distinguish them by source tbl.
.strategy

A strategy hint to provide to Kusto.
.shufflekeys

A character vector of column names to shuffle on, if .strategy = "shuffle".
.num_partitions

The number of partitions for a shuffle query.
.remote

A strategy hint to provide to Kusto for cross-cluster joins.

Append a set operation to the tbl_kusto object's ops list

Description

Append a set operation to the tbl_kusto object's ops list

Usage

add_op_set_op(x, y, type)

Arguments

x

The "left" tbl
y

The "right" tbl
type

The type of set operation to perform, currently only supports union_all

Append an operation representing a single-table verb to the tbl_kusto object's ops list

Description

Append an operation representing a single-table verb to the tbl_kusto object's ops list

Usage

add_op_single(name, .data, dots = list(), args = list())

Arguments

name

The name of the operation, e.g. 'select', 'filter'
.data

The tbl_kusto object to append the operation to
dots

The expressions passed as arguments to the operation verb
args

Other non-expression arguments passed to the operation verb

Kusto/Azure Data Explorer cluster resource class

Description

Class representing a Kusto cluster, exposing methods for working with it.

Methods

The following methods are available, in addition to those provided by the AzureRMR::az_resource class:

  • new(...): Initialize a new storage object. See 'Initialization'.

  • start(): Start the cluster.

  • stop(): Stop the cluster.

  • create_database(...): Create a new Kusto database. See Databases below.

  • ⁠get_database(database))⁠: Get an existing database.

  • delete_database(database, confirm=TRUE): Delete a database, by default asking for confirmation first.

  • list_databases(): List all databases in this cluster.

  • get_default_tenant(): Retrieve the default tenant to authenticate with this cluster.

  • get_query_token(tenant, ...): Obtain an authentication token from Azure Active Directory for this cluster's query endpoint. Accepts further arguments that will be passed to get_kusto_token.

  • get_ingestion_token(tenant, ...): Obtain an authentication token for this cluster's ingestion endpoint. Accepts further arguments that will be passed to get_kusto_token.

Initialization

Initializing a new object of this class can either retrieve an existing Kusto cluster, or create a new cluster on the host. Generally, the best way to initialize an object is via the get_kusto_cluster and create_kusto_cluster methods of the az_resource_group class, which handle the details automatically.

Databases

A Kusto cluster can have several databases, which are represented in AzureKusto via az_kusto_database R6 objects. The az_kusto class provides the create_database, get_database, delete_database and list_databases methods for creating, deleting and retrieving databases. It's recommended to use these methods rather than calling az_kusto_database$new() directly.

create_database takes the following arguments. It returns an object of class az_kusto_database

  • database: The name of the database to create.

  • retention_period: The retention period of the database, after which data will be soft-deleted.

  • cache_period: The cache period of the database, the length of time for which queries will be cached.

get_database takes a single argument database, the name of the database to retrieve, and returns an object of class az_kusto_database. delete_database takes the name of the database to delete and returns NULL on a successful deletion. list_databases takes no arguments and returns a list of az_kusto_database objects, one for each database in the cluster.

See Also

az_kusto_database, kusto_database_endpoint, create_kusto_cluster, get_kusto_cluster, delete_kusto_cluster, get_kusto_token

Kusto/Azure Data Explorer documentation,

Examples

## Not run: 

# recommended way of retrieving a resource: via a resource group object
kus <- resgroup$get_kusto_cluster("mykusto")

# list databases
kust$list_databases()

# create a new database with a retention period of 6 months
kust$create_database("newdb", retention_period=180)

# get the default authentication tenant
kus$get_default_tenant()

# generate an authentication token
kust$get_aad_token()


## End(Not run)

Kusto/Azure Data Explorer database resource class

Description

Class representing a Kusto database, exposing methods for working with it.

Methods

The following methods are available, in addition to those provided by the AzureRMR::az_resource class:

  • new(...): Initialize a new storage object. See 'Initialization'.

  • add_principals(...): Add new database principals. See Principals below.

  • remove_principals(...): Remove database principals.

  • list_principals(): Retrieve all database principals, as a data frame.

  • get_query_endpoint(): Get a query endpoint object for interacting with the database.

  • get_ingestion_endpoint(): Get an ingestion endpoint object for interacting with the database.

Initialization

Initializing a new object of this class can either retrieve an existing Kusto database, or create a new database on the server. Generally, the best way to initialize an object is via the get_database, list_databases() and create_database methods of the az_kusto class, which handle the details automatically.

Principals

This class provides methods for managing the principals of a database.

add_principal takes the following arguments. It returns a data frame with one row per principal, containing the details for each principal.

  • name: The name of the principal to create.

  • role: The role of the principal, for example "Admin" or "User".

  • type: The type of principal, either "User" or "App".

  • fqn: The fully qualified name of the principal, for example "aaduser=username@mydomain" for an Azure Active Directory account. If supplied, the other details will be obtained from this.

  • email: For a user principal, the email address.

  • app_id: For an application principal, the ID.

remove_principal removes a principal. It takes the same arguments as add_principal; if the supplied details do not match the actual details for the principal, it is not removed.

See Also

az_kusto, kusto_database_endpoint, create_database, get_database, delete_database

Kusto/Azure Data Explorer documentation,

Examples

## Not run: 

# recommended way of retrieving a resource: via a resource group object
db <- resgroup$
    get_kusto_cluster("mykusto")$
    get_database("mydatabase")

# list principals
db$list_principals()

# add a new principal
db$add_principal("New User", role="User", fqn="aaduser=username@mydomain")

# get the endpoint
db$get_database_endpoint(use_integer64=FALSE)


## End(Not run)

DBI interface: connect to a Kusto cluster

Description

Functions to connect to a Kusto cluster.

Usage

AzureKusto()

## S4 method for signature 'AzureKustoDriver'
dbConnect(drv, ..., bigint = c("numeric", "integer64"))

## S4 method for signature 'AzureKustoDriver'
dbCanConnect(drv, ...)

## S4 method for signature 'AzureKustoDriver'
dbDisconnect(conn, ...)

Arguments

drv

An AzureKusto DBI driver object, instantiated with AzureKusto().
...

Authentication arguments supplied to kusto_database_endpoint.
bigint

How to treat Kusto long integer columns. By default, they will be converted to R numeric variables. If this is "integer64", they will be converted to integer64 variables using the bit64 package.
conn

For dbDisconnect, an AzureKustoConnection object obtained with dbConnect.

Details

Kusto is connectionless, so dbConnect simply wraps a database endpoint object, generated with kusto_database_endpoint(...). The endpoint itself can be accessed via the ⁠@endpoint⁠ slot. Similarly, dbDisconnect always returns TRUE.

dbCanConnect attempts to detect whether querying the database with the given information and credentials will be successful. The result may not be accurate; essentially all it does is check that its arguments are valid Kusto properties. Ultimately the best way to tell if querying will work is to try it.

Value

For dbConnect, an object of class AzureKustoConnection.

For dbCanConnect, TRUE if authenticating with the Kusto server succeeded with the given arguments, and FALSE otherwise.

For dbDisconnect, always TRUE, invisibly.

See Also

kusto-DBI, dbReadTable, dbWriteTable, dbGetQuery, dbSendStatement, kusto_database_endpoint

Examples

## Not run: 
db <- DBI::dbConnect(AzureKusto(),
    server="https://mycluster.westus.kusto.windows.net", database="database", tenantid="contoso")

DBI::dbDisconnect(db)

# no authentication credentials: returns FALSE
DBI::dbCanConnect(AzureKusto(),
    server="https://mycluster.westus.kusto.windows.net")


## End(Not run)

Aggregation function translations

Description

Aggregation function translations

Usage

base_agg

Format

An object of class environment of length 7.

Scalar operator translations (infix and prefix)

Description

Scalar operator translations (infix and prefix)

Usage

base_scalar

Format

An object of class environment of length 79.

Window function translations

Description

Window function translations

Usage

base_window

Format

An object of class environment of length 1.

Build a KQL string.

Description

Build a KQL string.

Usage

build_kql(..., .env = parent.frame())

Arguments

...

input to convert to KQL. Use kql() to preserve user input as is (dangerous), and ident() to label user input as kql identifiers (safe)
.env

the environment in which to evaluate the arguments. Should not be needed in typical use.

Compile the preceding dplyr operations into a kusto query, execute it on the remote server, and return the result as a tibble.

Description

Compile the preceding dplyr operations into a kusto query, execute it on the remote server, and return the result as a tibble.

Usage

## S3 method for class 'tbl_kusto'
collect(x, ...)

Arguments

x

An instance of class tbl_kusto representing a Kusto table
...

needed for agreement with generic. Not otherwise used.

Execute the query, store the results in a table, and return a reference to the new table

Description

Execute the query, store the results in a table, and return a reference to the new table

Usage

## S3 method for class 'tbl_kusto'
compute(x, ..., name = generate_table_name())

Arguments

x

An instance of class tbl_kusto representing a Kusto table
...

other parameters passed to the query
name

The name for the Kusto table to be created. If name is omitted, the table will be named Rtbl_ + 8 random lowercase letters

This function uploads a local data frame into a remote data source, creating the table definition as needed. If the table exists, it will append the data to the existing table. If not, it will create a new table.

Description

This function uploads a local data frame into a remote data source, creating the table definition as needed. If the table exists, it will append the data to the existing table. If not, it will create a new table.

Usage

## S3 method for class 'kusto_database_endpoint'
copy_to(
  dest,
  df,
  name = deparse(substitute(df)),
  overwrite = FALSE,
  method = "inline",
  ...
)

Arguments

dest

remote data source
df

local data frame
name

Name for new remote table
overwrite

If TRUE, will overwrite an existing table with name name. If FALSE, will throw an error if name already exists.
method

For local ingestion, the method to use. "inline", "streaming", or "indirect".
...

other parameters passed to the query

See Also

collect() for the opposite action; downloading remote data into a local tbl.

Create Kusto/Azure Data Explorer cluster

Description

Method for the AzureRMR::az_resource_group class.

Usage

create_kusto_cluster(name, location,
                     node_size="D14_v2", ...)

Arguments

  • name: The name of the cluster.

  • location: The location/region in which to create the account. Defaults to the resource group location.

  • node_size: The capacity of the nodes in each of the cluster. Defaults to "D14_v2", which should be available in all regions. The availability of other sizes depends on the region the cluster is created in.

  • ... Other named arguments to pass to the az_kusto initialization function.

Details

This method deploys a new Kusto cluster resource, with parameters given by the arguments.

Value

An object of class az_kusto representing the created cluster.

See Also

get_kusto_cluster, delete_kusto_cluster, az_kusto

Kusto/Azure Data Explorer documentation

Examples

## Not run: 

rg <- AzureRMR::get_azure_login("myaadtenant")$
    get_subscription("subscription_id")$
    get_resource_group("rgname")

# create a new Kusto cluster
rg$create_kusto_cluster("mykusto", node_size="L16")


## End(Not run)

DBI methods for Kusto queries and commands

Description

DBI methods for Kusto queries and commands

Usage

## S4 method for signature 'AzureKustoConnection,character'
dbGetQuery(conn, statement, ...)

## S4 method for signature 'AzureKustoConnection'
dbSendQuery(conn, statement, ...)

## S4 method for signature 'AzureKustoResult'
dbFetch(res, n = -1, ...)

## S4 method for signature 'AzureKustoConnection,character'
dbSendStatement(conn, statement, ...)

## S4 method for signature 'AzureKustoConnection,character'
dbExecute(conn, statement, ...)

## S4 method for signature 'AzureKustoConnection,character'
dbListFields(conn, name, ...)

## S4 method for signature 'AzureKustoResult'
dbColumnInfo(res, ...)

Arguments

conn

An AzureKustoConnection object.
statement

A string containing a Kusto query or control command.
...

Further arguments passed to run_query.
res

An AzureKustoResult resultset object
n

The number of rows to return. Not used.
name

For dbListFields, a table name.

Details

These are the basic DBI functions to query the database. Note that Kusto only supports synchronous queries and commands; in particular, dbSendQuery and dbSendStatement will wait for the query or statement to complete, rather than returning immediately.

dbSendStatement and dbExecute are meant for running Kusto control commands, and will throw an error if passed a regular query. dbExecute also returns the entire result of running the command, rather than simply a row count.

See Also

dbConnect, dbReadTable, dbWriteTable, run_query

Examples

## Not run: 

db <- DBI::dbConnect(AzureKusto(),
    server="https://mycluster.location.kusto.windows.net", database="database"...)

DBI::dbGetQuery(db, "iris | count")
DBI::dbListFields(db, "iris")

# does the same thing as dbGetQuery, but returns an AzureKustoResult object
res <- DBI::dbSendQuery(db, "iris | count")
DBI::dbFetch(res)
DBI::dbColumnInfo(res)

DBI::dbExecute(db, ".show tables")

# does the same thing as dbExecute, but returns an AzureKustoResult object
res <- DBI::dbSendStatement(db, ".show tables")
DBI::dbFetch(res)


## End(Not run)

DBI methods for Kusto table management

Description

DBI methods for Kusto table management

Usage

## S4 method for signature 'AzureKustoConnection,character'
dbReadTable(conn, name, ...)

## S4 method for signature 'AzureKustoConnection,ANY'
dbWriteTable(conn, name, value, method, ...)

## S4 method for signature 'AzureKustoConnection'
dbCreateTable(conn, name, fields, ..., row.names = NULL, temporary = FALSE)

## S4 method for signature 'AzureKustoConnection,ANY'
dbRemoveTable(conn, name, ...)

## S4 method for signature 'AzureKustoConnection'
dbListTables(conn, ...)

## S4 method for signature 'AzureKustoConnection,ANY'
dbExistsTable(conn, name, ...)

Arguments

conn

An AzureKustoConnection object.
name

A string containing a table name.
...

Further arguments passed to run_query.
value

For dbWriteTable, a data frame to be written to a Kusto table.
method

For dbWriteTable, the ingestion method to use to write the table. See ingest_local.
fields

For dbCreateTable, the table specification: either a named character vector, or a data frame of sample values.
row.names

For dbCreateTable, the row names. Not used.
temporary

For dbCreateTable, whether to create a temporary table. Must be FALSE for Kusto.

Details

These functions read, write, create and delete a table, list the tables in a Kusto database, and check for table existence. With the exception of dbWriteTable, they ultimately call run_query which does the actual work of communicating with the Kusto server. dbWriteTable calls ingest_local to write the data to the server; note that it only supports ingesting a local data frame, as per the DBI spec.

Kusto does not have the concept of temporary tables, so calling dbCreateTable with temporary set to anything other than FALSE will generate an error.

dbReadTable and dbWriteTable are likely to be of limited use in practical scenarios, since Kusto tables tend to be much larger than available memory.

Value

For dbReadTable, an in-memory data frame containing the table.

See Also

AzureKusto-connection, dbConnect, run_query, ingest_local

Examples

## Not run: 
db <- DBI::dbConnect(AzureKusto(),
    server="https://mycluster.location.kusto.windows.net", database="database"...)

DBI::dbListTables(db)

if(!DBI::dbExistsTable(db, "mtcars"))
    DBI::dbCreateTable(db, "mtcars")

DBI::dbWriteTable(db, "mtcars", mtcars, method="inline")

DBI::dbReadTable(db, "mtcars")

DBI::dbRemoveTable(db, "mtcars")


## End(Not run)

Delete Kusto/Azure Data Explorer cluster

Description

Method for the AzureRMR::az_resource_group class.

Usage

delete_kusto_cluster(name, confirm=TRUE, wait=FALSE)

Arguments

  • name: The name of the cluster.

  • confirm: Whether to ask for confirmation before deleting.

  • wait: Whether to wait until the deletion is complete.

Value

NULL on successful deletion.

See Also

create_kusto_cluster, get_kusto_cluster, az_kusto

Kusto/Azure Data Explorer documentation

Examples

## Not run: 

rg <- AzureRMR::az_rm$
    new(tenant="myaadtenant.onmicrosoft.com", app="app_id", password="password")$
    get_subscription("subscription_id")$
    get_resource_group("rgname")

# delete a Kusto cluster
rg$delete_kusto_cluster("mycluster")


## End(Not run)

Escape/quote a string.

Description

Escape/quote a string.

Usage

escape(x, parens = NA, collapse = " ")

kql_vector(x, parens = NA, collapse = " ")

Arguments

x

An object to escape. Existing kql vectors will be left as is, character vectors are escaped with single quotes, numeric vectors have trailing .0 added if they're whole numbers, identifiers are escaped with double quotes.
parens, collapse

Controls behaviour when multiple values are supplied. parens should be a logical flag, or if NA, will wrap in parens if length > 1.

Default behaviour: lists are always wrapped in parens and separated by commas, identifiers are separated by commas and never wrapped, atomic vectors are separated by spaces and wrapped in parens if needed.

Execute the Kusto query and export the result to Azure Storage.

Description

Execute the Kusto query and export the result to Azure Storage.

Execute the Kusto query and export the result to Azure Storage.

Usage

export(
  tbl,
  storage_uri,
  query = NULL,
  name_prefix = "export",
  key = "impersonate",
  format = "parquet",
  distributed = FALSE,
  ...
)

## S3 method for class 'kusto_database_endpoint'
export(
  tbl,
  storage_uri,
  query = NULL,
  name_prefix = "export",
  key = "impersonate",
  format = "parquet",
  distributed = FALSE,
  ...
)

## S3 method for class 'tbl_kusto'
export(
  tbl,
  storage_uri,
  query = NULL,
  name_prefix = "export",
  key = "impersonate",
  format = "parquet",
  distributed = FALSE,
  ...
)

Arguments

tbl

A Kusto database endpoint object, as returned by kusto_database_endpoint.
storage_uri

The Azure Storage URI to export files to.
query

A Kusto query string
name_prefix

The filename prefix to use for exported files.
key

default "impersonate" which uses the account signed into Kusto to authenticate to Azure Storage. An Azure Storage account key.
format

Options are "parquet", "csv", "tsv", "json"
distributed

logical, indicates whether Kusto should distributed the export job to multiple nodes, in which case multiple files will be written to storage concurrently.
...

needed for agreement with generic. Not otherwise used.

Walks the tree of ops and builds a stack.

Description

Walks the tree of ops and builds a stack.

Usage

flatten_query(op, ops = list())

Arguments

op

the current operation
ops

the stack of operations to append to, recursively

Get existing Kusto/Azure Data Explorer cluster

Description

Method for the AzureRMR::az_resource_group class.

Usage

get_kusto_cluster(name, location,
                     node_size="D14_v2")

Arguments

  • name: The name of the cluster.

Details

This method retrieves an existing Kusto cluster resource.

Value

An object of class az_kusto representing the created cluster.

See Also

create_kusto_cluster, delete_kusto_cluster, az_kusto

Kusto/Azure Data Explorer documentation

Examples

## Not run: 

rg <- AzureRMR::get_azure_login("myaadtenant")$
    get_subscription("subscription_id")$
    get_resource_group("rgname")

# get a Kusto cluster
rg$get_kusto_cluster("mykusto")


## End(Not run)

Manage AAD authentication tokens for Kusto clusters

Description

Manage AAD authentication tokens for Kusto clusters

Usage

get_kusto_token(
  server = NULL,
  clustername,
  location = NULL,
  tenant = NULL,
  app = .kusto_app_id,
  auth_type = NULL,
  version = 2,
  ...
)

delete_kusto_token(
  server = NULL,
  clustername,
  location = NULL,
  tenant = NULL,
  app = .kusto_app_id,
  auth_type = NULL,
  version = 2,
  ...,
  hash = NULL,
  confirm = TRUE
)

list_kusto_tokens()

Arguments

server

The URI of your Kusto cluster. If not supplied, it is obtained from the clustername and location arguments.
clustername

The cluster name.
location

The cluster location. Leave this blank for a Microsoft-internal Kusto cluster like "help".
tenant

Your Azure Active Directory (AAD) tenant. Can be a GUID, a name ("myaadtenant") or a fully qualified domain name ("myaadtenant.com").
app

The ID of the Azure Active Directory app/service principal to authenticate with. Defaults to the ID of the KustoClient app.
auth_type

The authentication method to use. Can be one of "authorization_code", "device_code", "client_credentials" or "resource_owner". The default is to pick one based on the other arguments.
version

The AAD version to use. There should be no reason to change this from the default value of 2.
...

Other arguments to pass to AzureAuth::get_azure_token.
hash

For delete_kusto_token, the MD5 hash of the token. This is used to identify the token if provided.
confirm

For delete_kusto_token, whether to ask for confirmation before deleting the token.

Details

get_kusto_token returns an authentication token for the given cluster, caching its value on disk. delete_kusto_token deletes a cached token, and list_kusto_tokens lists all cached tokens.

By default, authentication tokens will be obtained using the main KustoClient Active Directory app. This app can be used to authenticate with any Kusto cluster (assuming, of course, you have the proper credentials).

Value

get_kusto_token returns an object of class AzureAuth::AzureToken representing the authentication token, while list_kusto_tokens returns a list of such objects. delete_azure_token returns NULL on a successful delete.

See Also

kusto_database_endpoint, AzureAuth::get_azure_token

Examples

## Not run: 

get_kusto_token("https://myclust.australiaeast.kusto.windows.net")
get_kusto_token(clustername="myclust", location="australiaeast")

# authenticate using client_credentials method: see ?AzureAuth::get_azure_token
get_kusto_token("https://myclust.australiaeast.kusto.windows.net",
                tenant="mytenant", app="myapp", password="password")


## End(Not run)

Flag a character string as a Kusto identifier

Description

Flag a character string as a Kusto identifier

Usage

ident(...)

Arguments

...

character strings to flag as Kusto identifiers

Pass an already-escaped string to Kusto

Description

Pass an already-escaped string to Kusto

Usage

ident_q(...)

Arguments

...

character strings to treat as already-escaped identifiers

Ingestion functions for Kusto

Description

Ingestion functions for Kusto

Usage

ingest_local(
  database,
  src,
  dest_table,
  method = NULL,
  staging_container = NULL,
  ingestion_token = database$token,
  http_status_handler = "stop",
  ...
)

ingest_url(database, src, dest_table, async = FALSE, ...)

ingest_blob(
  database,
  src,
  dest_table,
  async = FALSE,
  key = NULL,
  token = NULL,
  sas = NULL,
  ...
)

ingest_adls2(
  database,
  src,
  dest_table,
  async = FALSE,
  key = NULL,
  token = NULL,
  sas = NULL,
  ...
)

ingest_adls1(
  database,
  src,
  dest_table,
  async = FALSE,
  key = NULL,
  token = NULL,
  sas = NULL,
  ...
)

Arguments

database

A Kusto database endpoint object, created with kusto_database_endpoint.
src

The source data. This can be either a data frame, local filename, or URL.
dest_table

The name of the destination table.
method

For local ingestion, the method to use. See 'Details' below.
staging_container

For local ingestion, an Azure storage container to use for staging the dataset. This can be an object of class either AzureStor::blob_container or AzureStor::adls_filesystem. Only used if method="indirect".
ingestion_token

For local ingestion, an Azure Active Directory authentication token for the cluster ingestion endpoint. Only used if method="streaming".
http_status_handler

For local ingestion, how to handle HTTP conditions >= 300. Defaults to "stop"; alternatives are "warn", "message" and "pass". The last option will pass through the raw response object from the server unchanged, regardless of the status code. This is mostly useful for debugging purposes, or if you want to see what the Kusto REST API does. Only used if method="streaming".
...

Named arguments to be treated as ingestion parameters.
async

For the URL ingestion functions, whether to do the ingestion asychronously. If TRUE, the function will return immediately while the server handles the operation in the background.
key, token, sas

Authentication arguments for the Azure storage ingestion methods. If multiple arguments are supplied, a key takes priority over a token, which takes priority over a SAS. Note that these arguments are for authenticating with the Azure storage account, as opposed to Kusto itself.

Details

There are up to 3 possible ways to ingest a local dataset, specified by the method argument.

  • method="indirect": The data is uploaded to blob storage, and then ingested from there. This is the default if the AzureStor package is present.

  • method="streaming": The data is uploaded to the cluster ingestion endpoint. This is the default if the AzureStor package is not present, however be aware that currently (as of February 2019) streaming ingestion is in beta and has to be enabled for a cluster by filing a support ticket.

  • method="inline": The data is embedded into the command text itself. This is only recommended for testing purposes, or small datasets.

Note that the destination table must be created ahead of time for the ingestion to proceed.

Examples

## Not run: 

# ingesting from local:

# ingest via Azure storage
cont <- AzureStor::storage_container("https://mystorage.blob.core.windows.net/container",
    sas="mysas")
ingest_local(db, "file.csv", "table",
    method="indirect", storage_container=cont)

ingest_local(db, "file.csv", "table", method="streaming")

# ingest by inlining data into query
ingest_inline(db, "file.csv", "table", method="inline")

# ingesting online data:

# a public dataset: Microsoft web data from UCI machine learning repository
ingest_url(db,
    "https://archive.ics.uci.edu/ml/machine-learning-databases/anonymous/anonymous-msweb.data",
    "table")

# from blob storage:
ingest_blob(db,
    "https://mystorage.blob.core.windows.net/container/myblob",
    "table",
    sas="mysas")

# from ADLSGen2:
token <- AzureRMR::get_azure_token("https://storage.azure.com", "mytenant", "myapp", "password")
ingest_blob(db,
    "abfss://[email protected]/data/myfile",
    "table",
    token=token)


## End(Not run)

Join methods for Kusto tables

Description

These methods are the same as other joining methods, with the exception of the .strategy, .shufflekeys and .num_partitions optional arguments. They provide hints to the Kusto engine on how to execute the join, and can sometimes be useful to speed up a query. See the Kusto documentation for more details.

Usage

## S3 method for class 'tbl_kusto_abstract'
inner_join(
  x,
  y,
  by = NULL,
  copy = NULL,
  suffix = c(".x", ".y"),
  ...,
  keep = NULL,
  .strategy = NULL,
  .shufflekeys = NULL,
  .num_partitions = NULL,
  .remote = NULL
)

## S3 method for class 'tbl_kusto_abstract'
left_join(
  x,
  y,
  by = NULL,
  copy = NULL,
  suffix = c(".x", ".y"),
  ...,
  keep = NULL,
  .strategy = NULL,
  .shufflekeys = NULL,
  .num_partitions = NULL,
  .remote = NULL
)

## S3 method for class 'tbl_kusto_abstract'
right_join(
  x,
  y,
  by = NULL,
  copy = NULL,
  suffix = c(".x", ".y"),
  ...,
  keep = NULL,
  .strategy = NULL,
  .shufflekeys = NULL,
  .num_partitions = NULL,
  .remote = NULL
)

## S3 method for class 'tbl_kusto_abstract'
full_join(
  x,
  y,
  by = NULL,
  copy = NULL,
  suffix = c(".x", ".y"),
  ...,
  keep = NULL,
  .strategy = NULL,
  .shufflekeys = NULL,
  .num_partitions = NULL,
  .remote = NULL
)

## S3 method for class 'tbl_kusto_abstract'
semi_join(
  x,
  y,
  by = NULL,
  copy = NULL,
  ...,
  suffix = c(".x", ".y"),
  .strategy = NULL,
  .shufflekeys = NULL,
  .num_partitions = NULL,
  .remote = NULL
)

## S3 method for class 'tbl_kusto_abstract'
anti_join(
  x,
  y,
  by = NULL,
  copy = NULL,
  suffix = c(".x", ".y"),
  .strategy = NULL,
  .shufflekeys = NULL,
  .num_partitions = NULL,
  .remote = NULL,
  ...
)

Arguments

x, y

Kusto tbls.
by

The columns to join on.
copy

Needed for agreement with generic. Not otherwise used.
suffix

The suffixes to use for deduplicating column names.
...

Other arguments passed to lower-level functions.
keep

Needed for agreement with generic. Not otherwise used. Kusto retains keys from both sides of joins.
.strategy

A join strategy hint to pass to Kusto. Currently the values supported are "shuffle" and "broadcast".
.shufflekeys

A character vector of column names to use as shuffle keys.
.num_partitions

The number of partitions for a shuffle query.
.remote

A join strategy hint to use for cross-cluster joins. Can be "left", "right", "local" or "auto" (the default).

See Also

dplyr::join

Examples

## Not run: 

tbl1 <- tbl_kusto(db, "table1")
tbl2 <- tbl_kusto(db, "table2")

# standard dplyr syntax:
left_join(tbl1, tbl2)

# Kusto extensions:
left_join(tbl1, tbl2, .strategy = "broadcast") # a broadcast join

left_join(tbl1, tbl2, .shufflekeys = c("var1", "var2")) # shuffle join with shuffle keys

left_join(tbl1, tbl2, .num_partitions = 5) # no. of partitions for a shuffle join

## End(Not run)

Information functions

Description

These functions test whether an object is of the given class.

Usage

is_kusto_database(x)

is_kusto_cluster(x)

Arguments

x

An R object.

Tag character strings as Kusto Query Language. Assumes the string is valid and properly escaped.

Description

Tag character strings as Kusto Query Language. Assumes the string is valid and properly escaped.

Usage

kql(...)

Arguments

...

character strings to tag as KQL

Return a function representing a KQL aggregation function

Description

Return a function representing a KQL aggregation function

Usage

kql_aggregate(f)

Arguments

f

Name of the Kusto aggregation function

Build the tbl object into a data structure representing a Kusto query

Description

Build the tbl object into a data structure representing a Kusto query

Usage

kql_build(op)

Arguments

op

A nested sequence of query operations, i.e. tbl_kusto$ops

dplyr's mutate verb can include aggregations, but Kusto's extend verb cannot. If the mutate contains no aggregations, then it can emit an extend clause. If the mutate contains an aggregation and the tbl is ungrouped, then it must emit a summarize clause grouped by all variables. If the mutate contains an aggregation and the tbl is grouped, then it must join to a subquery containing the summarize clause.

Description

dplyr's mutate verb can include aggregations, but Kusto's extend verb cannot. If the mutate contains no aggregations, then it can emit an extend clause. If the mutate contains an aggregation and the tbl is ungrouped, then it must emit a summarize clause grouped by all variables. If the mutate contains an aggregation and the tbl is grouped, then it must join to a subquery containing the summarize clause.

Usage

## S3 method for class 'op_mutate'
kql_build(op, ...)

Arguments

op

A nested sequence of query operations, i.e. tbl_kusto$ops
...

Needed for agreement with generic. Not otherwise used.

Escape a Kusto identifier with [' ']

Description

Escape a Kusto identifier with [' ']

Usage

kql_escape_ident(x)

Arguments

x

An identifier to escape

Pass through an already-escaped Kusto identifier

Description

Pass through an already-escaped Kusto identifier

Usage

kql_escape_ident_q(x)

Arguments

x

An identifier to pass through

Escape a Kusto logical value. Converts TRUE/FALSE to true / false

Description

Escape a Kusto logical value. Converts TRUE/FALSE to true / false

Usage

kql_escape_logical(x)

Arguments

x

A logical value to escape

Escape a Kusto string by single-quoting

Description

Escape a Kusto string by single-quoting

Usage

kql_escape_string(x)

Arguments

x

A string to escape

Return a function representing a scalar KQL infix operator

Description

Return a function representing a scalar KQL infix operator

Usage

kql_infix(f)

Arguments

f

Name of a Kusto infix operator / function

Return a function representing a scalar KQL prefix function

Description

Return a function representing a scalar KQL prefix function

Usage

kql_prefix(f, n = NULL)

Arguments

f

Name of a Kusto infix function
n

Number of arguments accepted by the Kusto prefix function

Render a set of operations on a tbl_kusto_abstract to a Kusto query

Description

Render a set of operations on a tbl_kusto_abstract to a Kusto query

Usage

kql_render(query, ...)

Arguments

query

The tbl_kusto instance with a sequence of operations in $ops
...

needed for agreement with generic. Not otherwise used.

Build a kql_variant class out of the environments holding scalar and aggregation function definitions

Description

Build a kql_variant class out of the environments holding scalar and aggregation function definitions

Usage

kql_translate_env()

Builds an environment from a list of R -> Kusto query language translation pairs.

Description

Builds an environment from a list of R -> Kusto query language translation pairs.

Usage

kql_translator(..., .funs = list(), .parent = new.env(parent = emptyenv()))

Arguments

...

Pairs of R call = Kusto call translations as individual arguments
.funs

Parse of R call = Kusto call translations in list format
.parent

A parent environment to attach this env onto

Return a function representing a KQL window function

Description

Return a function representing a KQL window function

Usage

kql_window(f)

Arguments

f

Name of the Kusto aggregation function

Endpoints for communicating with a Kusto database

Description

Endpoints for communicating with a Kusto database

Usage

kusto_database_endpoint(
  ...,
  .connection_string = NULL,
  .query_token = NULL,
  .use_integer64 = FALSE
)

Arguments

...

Named arguments which are the properties for the endpoint object. See 'Details' below for the properties that AzureKusto recognises.
.connection_string

An alternative way of specifying the properties, as a database connection string. Properties supplied here override those in ... if they overlap.
.query_token

Optionally, an Azure Active Directory (AAD) token to authenticate with. If this is supplied, it overrides other tokens specified in ... or in the connection string.
.use_integer64

For kusto_database_endpoint, whether to convert columns with Kusto long datatype into 64-bit integers in R, using the bit64 package. If FALSE, represent them as numeric instead.

Details

This is a list of properties recognised by kusto_database_endpoint, and their alternate names. Property names not in this list will generate an error. Note that not all properties that are recognised are currently supported by AzureKusto.

General properties:

  • server: The URI of the server, usually of the form 'https://clustername.location.kusto.windows.net'.

    • addr, address, network address, datasource, host

  • database: The database.

    • initialcatalog, dbname

  • tenantid: The AAD tenant name or ID to authenticate with.

    • authority

  • appclientid: The AAD app/service principal ID

    • applicationclientid

  • traceclientversion: The client version for tracing.

  • queryconsistency: The level of query consistency. Defaults to "weakconsistency".

  • response_dynamic_serialization: How to serialize dynamic responses.

  • response_dynamic_serialization_2: How to serialize dynamic responses.

User authentication properties:

  • password

  • user: The user name.

    • uid, userid

  • traceusername: The user name for tracing.

  • usertoken: The AAD token for user authentication.

    • usertoken, usrtoken

  • fed: Logical, whether federated authentication is enabled. Currently unsupported; if this is TRUE, kusto_database_endpoint will print a warning and ignore it.

    • federated security, federated, aadfed, aadfederatedsecurity

App authentication properties:

  • appkey: The secret key for the app.

    • applicationkey

  • traceappname: The AAD app for tracing.

  • apptoken: The AAD token for app authentication.

    • apptoken, applicationtoken

Currently, AzureKusto only supports authentication via Azure Active Directory. Authenticating with DSTS is planned for the future.

The way kusto_database_endpoint obtains an AAD token is as follows.

  1. If the .query_token argument is supplied, use it.

  2. Otherwise, if the usertoken property is supplied, use it.

  3. Otherwise, if the apptoken property is supplied, use it.

  4. Otherwise, if the appclientid property is supplied, use it to obtain a token:

    • With the user and pwd properties if available

    • Or with the appkey property if available

    • Otherwise do an interactive authentication and ask for the user credentials

  5. Otherwise, if no appclientid property is supplied, authenticate with the KustoClient app:

    • With the user and pwd properties if available

    • Otherwise do an interactive authentication and ask for the user credentials using a device code

Value

An object of class kusto_database_endpoint.

See Also

run_query, az_kusto_database

Examples

## Not run: 

kusto_database_endpoint(server="myclust.australiaeast.kusto.windows.net", database="db1")

# supplying a token obtained previously
token <- get_kusto_token("myclust.australiaeast.kusto.windows.net")
kusto_database_endpoint(server="myclust.australiaeast.kusto.windows.net", database="db1",
                        .query_token=token)


## End(Not run)

Execute the query, store the results in a table, and return a reference to the new table Run a Kusto query and export results to Azure Storage in Parquet or CSV format.

Description

Execute the query, store the results in a table, and return a reference to the new table Run a Kusto query and export results to Azure Storage in Parquet or CSV format.

Usage

kusto_export_cmd(query, storage_uri, name_prefix, key, format, distributed)

Arguments

query

The text of the Kusto query to run
storage_uri

The URI of the blob storage container to export to
name_prefix

The filename prefix for each exported file
key

The account key for the storage container. uses the identity that is signed into Kusto to authenticate to Azure Storage.
format

Options are "parquet", "csv", "tsv", "json"
distributed

logical, indicates whether Kusto should distributed the export job to multiple nodes, in which case multiple files will be written to storage concurrently.

DBI interface to Kusto

Description

AzureKusto implements a subset of the DBI specification for interfacing with databases in R. The following methods are supported:

Details

Kusto is quite different to the SQL databases that DBI targets, which affects the behaviour of certain DBI methods and renders other moot.

  • Kusto is connectionless. dbConnect simply wraps a database endpoint object, created with kusto_database_endpoint. Similarly, dbDisconnect always returns TRUE. dbCanConnect attempts to check if querying the database will succeed, but this may not be accurate.

  • Temporary tables are not a Kusto concept, so ⁠dbCreateTable(*, temporary=TRUE)⁠ will throw an error.

  • It only supports synchronous queries, with a default timeout of 4 minutes. dbSendQuery and dbSendStatement will wait for the query to execute, rather than returning immediately. The object returned contains the full result of the query, which dbFetch extracts.

  • The Kusto Query Language (KQL) is not SQL, and so higher-level SQL methods are not implemented.

Nest method for Kusto tables

Description

This method collapses a column into a list

Usage

## S3 method for class 'tbl_kusto_abstract'
nest(.data, ...)

Arguments

.data

A kusto tbl.
...

Specification of columns to nest. Translates to summarize make_list() in Kusto.

The "base case" operation representing the tbl itself and its column variables

Description

The "base case" operation representing the tbl itself and its column variables

Usage

op_base(x, vars, class = character())

Arguments

x

A tbl object
vars

A vector of column variables in the tbl
class

The class that op_base should inherit from, default is character()

A double-table verb, e.g. joins, setops

Description

A double-table verb, e.g. joins, setops

Usage

op_double(name, x, y, args = list())

Arguments

name

The name of the operation, e.g. 'left_join', 'union_all'
x

The "left" tbl
y

The "right" tbl
args

Other arguments passed to the operation verb

Look up the applicable grouping variables for an operation based on the data source and preceding sequence of operations

Description

Look up the applicable grouping variables for an operation based on the data source and preceding sequence of operations

Usage

op_grps(op)

Arguments

op

An operation instance

A class representing a single-table verb

Description

A class representing a single-table verb

Usage

op_single(name, x, dots = list(), args = list())

Arguments

name

the name of the operation verb, e.g. "select", "filter"
x

the tbl object
dots

expressions passed to the operation verb function
args

other arguments passed to the operation verb function

Look up the applicable variables in scope for a given operation based on the data source and preceding sequence of operations

Description

Look up the applicable variables in scope for a given operation based on the data source and preceding sequence of operations

Usage

op_vars(op)

Arguments

op

An operation instance

Run a query or command against a Kusto database

Description

Run a query or command against a Kusto database

Usage

run_query(database, qry_cmd, ..., .http_status_handler = "stop")

Arguments

database

A Kusto database endpoint object, as returned by kusto_database_endpoint.
qry_cmd

A string containing the query or command. In KQL, a database management command is a statement that starts with a "."
...

Named arguments to be used as parameters for a parameterized query. These are ignored for database management commands.
.http_status_handler

The function to use to handle HTTP status codes. The default "stop" will throw an R error via httr::stop_for_status if the status code is not less than 300; other possibilities are "warn", "message" and "pass". The last option will pass through the raw response object from the server unchanged, regardless of the status code. This is mostly useful for debugging purposes, or if you want to see what the Kusto REST API does.

Details

This function is the workhorse of the AzureKusto package. It communicates with the Kusto server and returns the query or command results, as data frames.

See Also

kusto_database_endpoint, ingest_local, ingest_url, ingest_blob, ingest_adls2

Examples

## Not run: 

endp <- kusto_database_endpoint(server="myclust.australiaeast.kusto.windows.net", database="db1")

# a command
run_query(endp, ".show table iris")

# a query
run_query(endp, "iris | count")


## End(Not run)

Translate a sequence of dplyr operations on a tbl into a Kusto query string.

Description

Translate a sequence of dplyr operations on a tbl into a Kusto query string.

Usage

## S3 method for class 'tbl_kusto_abstract'
show_query(x, ...)

Arguments

x

A tbl_kusto or tbl_kusto_abstract instance
...

needed for agreement with generic. Not otherwise used.

Summarise method for Kusto tables

Description

This method is the same as other summarise methods, with the exception of the .strategy, .shufflekeys and .num_partitions optional arguments. They provide hints to the Kusto engine on how to execute the summarisation, and can sometimes be useful to speed up a query. See the Kusto documentation for more details.

Usage

## S3 method for class 'tbl_kusto_abstract'
summarise(
  .data,
  ...,
  .strategy = NULL,
  .shufflekeys = NULL,
  .num_partitions = NULL
)

Arguments

.data

A Kusto tbl.
...

Summarise expressions.
.strategy

A summarise strategy to pass to Kusto. Currently the only value supported is "shuffle".
.shufflekeys

A character vector of column names to use as shuffle keys.
.num_partitions

The number of partitions for a shuffle query.

See Also

dplyr::summarise

Examples

## Not run: 

tbl1 <- tbl_kusto(db, "table1")

## standard dplyr syntax:
summarise(tbl1, mx = mean(x))

## Kusto extensions:
summarise(tbl1, mx = mean(x), .strategy = "broadcast") # a broadcast summarise

summarise(tbl1, mx = mean(x), .shufflekeys = c("var1", "var2")) # a shuffle summarise with keys

summarise(tbl1, mx = mean(x), .num_partitions = 5) # no. of partitions for a shuffle summarise

## End(Not run)

A tbl object representing a table in a Kusto database.

Description

A tbl object representing a table in a Kusto database.

Usage

tbl_kusto(kusto_database, table_name, ...)

Arguments

kusto_database

An instance of kusto_database_endpoint that this table should be queried from
table_name

The name of the table in the Kusto database
...

parameters to pass in case the Kusto source table is a parameterized function.

Translate R expressions into Kusto Query Language equivalents.

Description

Translate R expressions into Kusto Query Language equivalents.

Usage

translate_kql(...)

Arguments

...

Expressions to translate.

Unnest method for Kusto tables

Description

This method takes a list column and expands it so that each element of the list gets its own row. unnest() translates to Kusto's mv-expand operator.

Usage

## S3 method for class 'tbl_kusto_abstract'
unnest(
  data,
  cols,
  ...,
  keep_empty = FALSE,
  ptype = NULL,
  names_sep = NULL,
  names_repair = NULL,
  .drop = NULL,
  .id = NULL,
  .sep = NULL,
  .preserve = NULL
)

Arguments

data

A Kusto tbl.
cols

Specification of columns to unnest.
...

[Deprecated]: previously you could write df %>% unnest(x, y, z). Convert to df %>% unnest(c(x, y, z)). If you previously created a new variable in unnest() you'll now need to do it explicitly with mutate(). Convert df %>% unnest(y = fun(x, y, z)) to df %>% mutate(y = fun(x, y, z)) %>% unnest(y).
keep_empty

Needed for agreement with generic. Not otherwise used. Kusto does not keep empty rows.
ptype

Needed for agreement with generic. Not otherwise used.
names_sep

Needed for agreement with generic. Not otherwise used.
names_repair

Needed for agreement with generic. Not otherwise used.
.drop

Needed for agreement with generic. Not otherwise used.
.id

Data frame identifier - if supplied, will create a new column with name .id, giving a unique identifier. This is most useful if the list column is named.
.sep

Needed for agreement with generic. Not otherwise used.
.preserve

Needed for agreement with generic. Not otherwise used.