In ropensci/nodbi: 'NoSQL' Database Connector

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

Summary

R package nodbi provides a single interface for several NoSQL databases and SQL databases with JSON functionality, with the same function parameters and return values across all of:

• MongoDB
• SQLite
• PostgreSQL
• DuckDB
• Elasticsearch
• CouchDB

Package nodbi has been designed to use any specific SQL functions a database may have for JSON and has added functionality tested for performance to enable switching databases without changing user code.

library(nodbi)

Functionality

Connect

First, a connection to a database is opened. In the example, no additional parameters are used such as database file or server; see the help page for the respective database.

"Container" is used as term to indicate where conceptually the database holds the data (e.g. a collection in MongoDB, a table in DuckDB). The key parameter of nodbi functions holds the name of the relevant container.

# name of container
key <- "my_container"

# nodbi can connect any of these databases
if (FALSE) {
  src <- src_duckdb()
  src <- src_mongo(collection = key)
  src <- src_sqlite()
  src <- src_postgres()
  src <- src_elastic()
  src <- src_couchdb(
    user = Sys.getenv("COUCHDB_TEST_USER"),
    pwd = Sys.getenv("COUCHDB_TEST_PWD")
  )
}

# this example is run with
src <- src_sqlite()

# note additional parameters can be specified,
# for example for local or remote MongoDb:
help("src_mongo")

docdb_create

Create a container if it does not yet exist and fill with value. The return value is the number of created documents. "Documents" refers to the rows in a data frame such as mtcars, or the number of NDJSON lines, or the number of list items, or the number of objects in an JSON array.

The parameter value in any nodbi function can take a data frame, a list, a JSON string, or a file name or URL pointing to NDJSON.

# check if container already exists
docdb_exists(src = src, key = key)

# load data from a data frame with row names into
# the container specified in "key" parameter
docdb_create(src = src, key = key, value = mtcars)

# load additionally 98 NDJSON records
docdb_create(src, key, "https://httpbin.org/stream/98")

# load additionally mapdata as list
docdb_create(src, key, jsonlite::fromJSON(mapdata, simplifyVector = FALSE))

# show JSON structure of contacts
jsonlite::minify(contacts)

# load additionally contacts JSON data
docdb_create(src, key, contacts)

Check and list any other containers exist in the database:

docdb_list(src = src)

Identifiers

The unique document identifier is its _id, corresponding to a primary index with a constraint to be unique in SQL databases.

The _id's of an input value are either the row names of a data frame (such as mtcars) or top-level elements with the name _id such as in contacts shown just above.

Thus, expect a warning when trying to create documents with _id's that already exist in the container.

The return value can be 0 when no documents could newly be created, or the number of the subset of documents in value that did not yet exist and could newly be created.

# zero new documents created
docdb_create(src, key, value = mtcars)

For updating existing documents, see below function docdb_update().

docdb_get

All documents in a container can now be retrieved with docdb_get().

# load library for more
# readable print output
library(tibble)

# get all documents, irrespective of schema
as_tibble(docdb_get(src, key))

# get just 2 documents using limit and note that
# only fields for these documents are returned
as_tibble(docdb_get(src, key, limit = 2L))

docdb_query

One of the most powerful functions of nodbi is docdb_query() because it permits to combine a query to select documents and to filter for fields of interest.

# query for some documents
docdb_query(src = src, key = key, query = '{"mpg": {"$gte": 30}}')

Both parameters query (obligatory) and fields (optional) use, across all databases, MongoDB syntax such as documented for queries and fields.

# query some fields from some documents; 'query' is a mandatory
# parameter and is used here in its position in the signature
docdb_query(src, key, '{"mpg": {"$gte": 30}}', fields = '{"wt": 1, "mpg": 1}')

Unless fields specifies "_id": 0, the _id field is always included in the output of docdb_query().

# query some fields from some documents, limit return to one document
docdb_query(src, key, '{"mpg": {"$gte": 30}}', fields = '{"_id": 0, "mpg": 1}', limit = 1L)

Queries can be more complex such as in this example, showing a dot notation of a sub-field and an example operator (regular expression).

# query some subitem fields from some documents
str(docdb_query(
  src, key,
  query = '{"$or": [{"age": {"$gt": 21}},
           {"friends.name": {"$regex": "^B[a-z]{3,9}.*"}}]}',
  fields = '{"age": 1, "friends.name": 1}'
))

Queries work across documents of different structure such as here.

# query with results across documents
docdb_query(
  src, key,
  query = '{"$or": [{"age": {"$gt": 21}}, {"mpg": {"$gte": 30}}]}',
  fields = '{"name": 1, "disp": 1}'
)

Field names

The JSON data handled by package nodbi may have a large number of field included nested fields in objects (see for example name within array friends above). Thus, an argument is provided for docdb_query() so that the function returns only the comprehensive list of all field names in documents selected with a query (or in all documents in the container if query = "{}" is specified).

docdb_query(src, key, query = '{"_id": {"$regex": "^[0-9]"}}', listfields = TRUE)

The dot notation is a path from a root field to the nested field, and this notation can be used in query and fields parameters of docdb_query().

docdb_update

Queries can also be used for updating (patching) selected documents with a new value. The return value of docdb_update() corresponds to the number of documents that were updated.

This is another powerful function because value can come from a data frame, a list, a JSON string, or a file name or URL pointing to NDJSON, and if value includes row names or _id's, these are used to identify the documents to be updated.

# number of documents corresponding to query
nrow(docdb_query(src, key, query = '{"carb": 3}'))

# update all documents using JSON, replacing the previously existing values
docdb_update(src, key, value = '{"vs": 9, "xy": [1, 2]}', query = '{"carb": 3}')

# update with value that includes _id's
docdb_update(src, key, value = '{"_id": "Merc 450SLC", "xy": 33}', query = "{}")

# show updated values
docdb_query(src, key, query = '{"carb": 3}', fields = '{"xy": 1}')

docdb_delete

Documents and containers can be deleted with docdb_delete(). Its return value corresponds to the success of the delete operation.

# number of documents corresponding to query
nrow(docdb_query(src, key, query = '{"age": {"$lte": 23}}'))

# to delete selected documents, specify a query parameter
docdb_delete(src, key, query = '{"age": {"$lte": 23}}')

# this deletes the complete container from database
docdb_delete(src, key)

# check if still exists
docdb_exists(src, key)

Disconnect and shutdown

Package nodbi includes an automatic mechanism for shutting down, at the time of quit() or session restart, those databases that require it (SQLite, DuckDB, PostgreSQL).

Nevertheless, it is good practice to manually disconnect and shut down connections as specific to the database, for example for SQLite:

src

# shutdown
DBI::dbDisconnect(src$con, shutdown = TRUE)
rm(src)

ropensci/nodbi documentation built on May 5, 2024, 5:42 p.m.