NAME
dsquery
SYNOPSIS
dsquery OPTIONS C_NAME SQL_STATEMENT
DESCRIPTION
dsquery is a tool to support SQL queries of dataset
collections. Pairtree based collections should be index before trying to
query them (see ‘-index’ option below). Pairtree collections use the
SQLite 3 dialect of SQL for querying. For collections using a SQL
storage engine (e.g. SQLite3 and Postgres), the SQL dialect reflects the
SQL of the storage engine.
The schema is the same for all storage engines. The scheme for the
JSON stored documents have a four column scheme. The columns are “_key”,
“created”, “updated” and “src”. “_key” is a string (aka VARCHAR),
“created” and “updated” are timestamps while “src” is a JSON column
holding the JSON document. The table name reflects the collection name
without the “.ds” extension (e.g. data.ds is stored in a database called
data having a table also called data).
The output of dsquery is a JSON array of objects.
The order of the objects is determined by the your SQL statement and SQL
engine. There is an option to generate a 2D grid of values in JSON, CSV
or YAML formats. See OPTIONS for details.
PARAMETERS
- C_NAME
-
If harvesting the dataset collection name to harvest the records to.
- SQL_STATEMENT
-
The SQL statement should conform to the SQL dialect used for the JSON
store(SQLite3 or Postgres). If it is a select statement is must return a
single column of type JSON, TEXT, VARCHAR, BOOLEAN, INTEGER, REAL. DATE,
DATESTAMP, TIME and TIMESTAMP columns are mapped to JSON string type.
dsquery returns an JSON array of value JSON objects or
types. If the query does not return a value (is not select statement)
than an empty JSON array is returned.
SQL Store Scheme
- _key
-
The key or id used to identify the JSON documented stored.
- src
-
This is a JSON column holding the JSON document
- created
-
The date the JSON document was created in the table
- updated
-
The date the JSON document was updated
OPTIONS
- -help
-
display help
- -license
-
display license
- -version
-
display version
- -pretty
-
pretty print the resulting JSON array
- -sql SQL_FILENAME
-
read SQL from a file. If filename is “-” then read SQL from standard
input.
- -grid STRING_OF_ATTRIBUTE_NAMES
-
Returns list as a 2D grid of values. This options requires a comma
delimited string of attribute names for the outer object to include in
grid output. It can be combined with -pretty options.
- -csv STRING_OF_ATTRIBUTE_NAMES
-
Like -grid this takes our list of dataset objects and a list of
attribute names but rather than create a 2D JSON array of values it
creates CSV representation with the first row as the attribute names.
- -yaml STRING_OF_ATTRIBUTE_NAMES
-
Like -grid this takes our list of dataset objects and a list of
attribute names but rather than create a 2D JSON of values it creates
YAML representation.
- -index
-
This will create a SQLite3 index for a collection. This enables dsquery
to query pairtree collections using SQLite3 SQL dialect just as it would
for SQL storage collections (i.e. don’t use with postgres or sqlite
based dataset collections. It is not needed for them). Note the index is
always built before executing the SQL statement.
EXAMPLES
Generate a list of JSON objects with the _key value
merged with the object stored as the ._Key attribute. The
colllection name “data.ds” which is implemented using Postgres as the
JSON store. (note: in Postgres the || is very helpful).
dsquery data.ds "SELECT jsonb_build_object('_Key', _key)::jsonb || src::jsonb FROM data"
In this example we’re returning the “src” in our collection by
querying for a “id” attribute in the “src” column. The id is passed in
as an attribute using the Postgres positional notatation in the
statement.
dsquery data.ds "SELECT src FROM data WHERE src->>'id' = $1 LIMIT 1" "xx103-3stt9"
