SQL Dialects

FlowScope supports 13 SQL dialects with varying levels of syntax coverage. Selecting the correct dialect ensures accurate parsing of dialect-specific functions, quoting styles, and statement syntax.

Supported Dialects

Dialect	Use For
generic	Unknown or mixed SQL
ansi	Standard ANSI SQL
postgresql	PostgreSQL, CockroachDB
snowflake	Snowflake Data Cloud
bigquery	Google BigQuery
duckdb	DuckDB
redshift	Amazon Redshift
mysql	MySQL, MariaDB
mssql	Microsoft SQL Server
sqlite	SQLite
hive	Apache Hive
databricks	Databricks SQL
clickhouse	ClickHouse

Selecting a Dialect

Web App

Use the dialect dropdown in the editor toolbar. The default is generic which handles most standard SQL.

CLI

Use the -d or --dialect flag:

flowscope -d snowflake queries/*.sql
flowscope --dialect bigquery analytics.sql

API

Pass the dialect in the request:

const result = await analyzeSql({
  sql: 'SELECT * FROM orders',
  dialect: 'postgresql'
});

Dialect-Specific Syntax

PostgreSQL

-- Array syntax
SELECT array_agg(name) FROM users;
SELECT tags[1] FROM posts;

-- Dollar quoting
SELECT $tag$literal string$tag$;

-- Type casting
SELECT '2024-01-01'::date;
SELECT CAST(amount AS numeric(10,2));

-- Array operations
SELECT * FROM orders WHERE tags @> ARRAY['urgent'];

Snowflake

-- Lateral flatten
SELECT f.value:name::string
FROM orders,
LATERAL FLATTEN(input => items) f;

-- Semi-structured data
SELECT data:customer:name FROM events;

-- Variant type
SELECT PARSE_JSON('{"key": "value"}');

-- Time travel
SELECT * FROM orders AT(TIMESTAMP => '2024-01-01'::timestamp);

BigQuery

-- Struct and array
SELECT STRUCT(1 AS a, 'hello' AS b);
SELECT ARRAY<INT64>[1, 2, 3];

-- Unnest
SELECT * FROM UNNEST([1, 2, 3]) AS num;

-- Safe navigation
SELECT SAFE.PARSE_DATE('%Y-%m-%d', date_str);

-- Backtick identifiers
SELECT * FROM `project.dataset.table`;

DuckDB

-- List comprehension
SELECT [x * 2 FOR x IN [1, 2, 3]];

-- Struct access
SELECT struct.field FROM data;

-- Positional joins
SELECT * FROM t1 POSITIONAL JOIN t2;

-- Friendly SQL
SELECT * FROM 'data.parquet';
SELECT * FROM read_csv_auto('file.csv');

MySQL

-- Backtick identifiers
SELECT `column` FROM `table`;

-- LIMIT syntax
SELECT * FROM orders LIMIT 10, 20;

-- String functions
SELECT CONCAT_WS('-', year, month, day);

-- INSERT ... ON DUPLICATE KEY
INSERT INTO t (id, val) VALUES (1, 'a')
ON DUPLICATE KEY UPDATE val = VALUES(val);

Redshift

-- DISTKEY and SORTKEY
CREATE TABLE orders (
    id INT,
    date DATE SORTKEY
) DISTSTYLE KEY DISTKEY(id);

-- COPY command
COPY orders FROM 's3://bucket/data'
CREDENTIALS 'aws_iam_role=...'
FORMAT AS PARQUET;

-- Spectrum external tables
SELECT * FROM spectrum.external_table;

Identifier Quoting

Different dialects use different quoting styles:

Dialect	Quote Style	Example
PostgreSQL	Double quotes	"Column Name"
MySQL	Backticks	`Column Name`
BigQuery	Backticks	`project.dataset.table`
SQL Server	Brackets	[Column Name]
ANSI	Double quotes	"Column Name"

FlowScope normalizes identifiers internally for lineage tracking.

Function Support

FlowScope recognizes dialect-specific functions:

Aggregations

Function	Dialects
ARRAY_AGG	PostgreSQL, BigQuery, Snowflake
STRING_AGG	PostgreSQL, BigQuery
GROUP_CONCAT	MySQL
LISTAGG	Snowflake, Redshift
LIST	DuckDB

Date Functions

Function	Dialects
DATE_TRUNC	PostgreSQL, Snowflake, BigQuery, DuckDB
DATE_PART	PostgreSQL, Snowflake, Redshift
EXTRACT	All dialects
DATEDIFF	Snowflake, MySQL, SQL Server
DATE_SUB	BigQuery, MySQL

String Functions

Function	Dialects
CONCAT	All dialects
SUBSTRING	ANSI standard
SUBSTR	Oracle-style dialects
SPLIT	BigQuery, Snowflake
REGEXP_EXTRACT	BigQuery
REGEXP_MATCHES	PostgreSQL

Limitations

Unsupported Syntax

Some dialect features produce parsing warnings:

• Procedural code (stored procedures, functions with bodies)
• Dialect-specific hints and pragmas
• Administrative commands (GRANT, REVOKE)
• Exotic statement types (EXPLAIN, ANALYZE)

FlowScope skips unsupported statements and continues with the rest.

Generic Dialect

When using the generic dialect:

• Only ANSI-standard syntax is guaranteed
• Dialect-specific functions may not be recognized
• Some valid SQL may produce parse errors

Always select the specific dialect when known for best results.

Dialect Detection

FlowScope does not auto-detect dialects. Common indicators to help you choose:

Indicator	Likely Dialect
Backtick identifiers	MySQL, BigQuery
:: type casts	PostgreSQL, DuckDB
FLATTEN	Snowflake
UNNEST	BigQuery
DISTKEY	Redshift
Dollar quoting	PostgreSQL