Skip to content

fix: preserve schema qualifiers in column default type casts #219

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
tianzhou merged 1 commit into from
Dec 30, 2025
Merged

fix: preserve schema qualifiers in column default type casts #219

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 2 additions & 1 deletion cmd/dump/dump_integration_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -256,7 +256,8 @@ func runTenantSchemaTest(t *testing.T, testDataDir string) {
}
}

// normalizeSchemaOutput removes version-specific lines for comparison
// normalizeSchemaOutput removes version-specific lines for comparison.
// This allows comparing dumps across different PostgreSQL versions.
func normalizeSchemaOutput(output string) string {
lines := strings.Split(output, "\n")
var normalizedLines []string
Expand Down
8 changes: 8 additions & 0 deletions ir/normalize.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -141,6 +141,14 @@ func normalizeDefaultValue(value string, tableSchema string) string {
re := regexp.MustCompile(`::pgschema_tmp_[^.]+\.`)
value = re.ReplaceAllString(value, "::")
}
// Also strip same-schema type qualifiers for consistent comparison during plan/diff
// This ensures that '::public.typename' from current state matches '::typename' from
// desired state (after pgschema_tmp_* is stripped). Both are semantically equivalent
// within the same schema context. (Issue #218)
if tableSchema != "" && strings.Contains(value, "::"+tableSchema+".") {
re := regexp.MustCompile(`::\Q` + tableSchema + `\E\.`)
value = re.ReplaceAllString(value, "::")
}

// Handle NULL::type -> NULL
// Example: NULL::text -> NULL
Expand Down
200 changes: 124 additions & 76 deletions ir/queries/queries.sql
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -56,74 +56,116 @@ ORDER BY t.table_name;

-- GetColumns retrieves all columns for all tables
-- name: GetColumns :many
WITH column_base AS (
SELECT
c.table_schema,
c.table_name,
c.column_name,
c.ordinal_position,
c.column_default,
c.is_nullable,
c.data_type,
c.character_maximum_length,
c.numeric_precision,
c.numeric_scale,
c.udt_name,
COALESCE(d.description, '') AS column_comment,
CASE
WHEN dt.typtype = 'd' THEN
CASE WHEN dn.nspname = c.table_schema THEN dt.typname
ELSE dn.nspname || '.' || dt.typname
END
WHEN dt.typtype = 'e' OR dt.typtype = 'c' THEN
CASE WHEN dn.nspname = c.table_schema THEN dt.typname
ELSE dn.nspname || '.' || dt.typname
END
WHEN dt.typtype = 'b' AND dt.typelem <> 0 THEN
-- Array types: apply same schema qualification logic to element type
CASE
WHEN en.nspname = 'pg_catalog' THEN et.typname || '[]'
WHEN en.nspname = c.table_schema THEN et.typname || '[]'
ELSE en.nspname || '.' || et.typname || '[]'
END
WHEN dt.typtype = 'b' THEN
-- Non-array base types: qualify if not in pg_catalog or table's schema
CASE
WHEN dn.nspname = 'pg_catalog' THEN c.udt_name
WHEN dn.nspname = c.table_schema THEN dt.typname
ELSE dn.nspname || '.' || dt.typname
END
ELSE c.udt_name
END AS resolved_type,
c.is_identity,
c.identity_generation,
c.identity_start,
c.identity_increment,
c.identity_maximum,
c.identity_minimum,
c.identity_cycle,
a.attgenerated,
ad.adbin,
ad.adrelid
FROM information_schema.columns c
LEFT JOIN pg_class cl ON cl.relname = c.table_name
LEFT JOIN pg_namespace n ON cl.relnamespace = n.oid AND n.nspname = c.table_schema
LEFT JOIN pg_description d ON d.objoid = cl.oid AND d.classoid = 'pg_class'::regclass AND d.objsubid = c.ordinal_position
LEFT JOIN pg_attribute a ON a.attrelid = cl.oid AND a.attname = c.column_name
LEFT JOIN pg_attrdef ad ON ad.adrelid = a.attrelid AND ad.adnum = a.attnum
LEFT JOIN pg_type dt ON dt.oid = a.atttypid
LEFT JOIN pg_namespace dn ON dt.typnamespace = dn.oid
LEFT JOIN pg_type et ON dt.typelem = et.oid
LEFT JOIN pg_namespace en ON et.typnamespace = en.oid
WHERE
c.table_schema NOT IN ('information_schema', 'pg_catalog', 'pg_toast')
AND c.table_schema NOT LIKE 'pg_temp_%'
AND c.table_schema NOT LIKE 'pg_toast_temp_%'
)
SELECT
c.table_schema,
c.table_name,
c.column_name,
c.ordinal_position,
CASE
WHEN a.attgenerated = 's' THEN NULL -- Generated columns don't have defaults
ELSE COALESCE(pg_get_expr(ad.adbin, ad.adrelid), c.column_default)
END AS column_default,
c.is_nullable,
c.data_type,
c.character_maximum_length,
c.numeric_precision,
c.numeric_scale,
c.udt_name,
COALESCE(d.description, '') AS column_comment,
CASE
WHEN dt.typtype = 'd' THEN
CASE WHEN dn.nspname = c.table_schema THEN dt.typname
ELSE dn.nspname || '.' || dt.typname
END
WHEN dt.typtype = 'e' OR dt.typtype = 'c' THEN
CASE WHEN dn.nspname = c.table_schema THEN dt.typname
ELSE dn.nspname || '.' || dt.typname
END
WHEN dt.typtype = 'b' AND dt.typelem <> 0 THEN
-- Array types: apply same schema qualification logic to element type
CASE
WHEN en.nspname = 'pg_catalog' THEN et.typname || '[]'
WHEN en.nspname = c.table_schema THEN et.typname || '[]'
ELSE en.nspname || '.' || et.typname || '[]'
END
WHEN dt.typtype = 'b' THEN
-- Non-array base types: qualify if not in pg_catalog or table's schema
CASE
WHEN dn.nspname = 'pg_catalog' THEN c.udt_name
WHEN dn.nspname = c.table_schema THEN dt.typname
ELSE dn.nspname || '.' || dt.typname
END
ELSE c.udt_name
END AS resolved_type,
c.is_identity,
c.identity_generation,
c.identity_start,
c.identity_increment,
c.identity_maximum,
c.identity_minimum,
c.identity_cycle,
a.attgenerated,
CASE
WHEN a.attgenerated = 's' THEN pg_get_expr(ad.adbin, ad.adrelid)
ELSE NULL
END AS generated_expr
FROM information_schema.columns c
LEFT JOIN pg_class cl ON cl.relname = c.table_name
LEFT JOIN pg_namespace n ON cl.relnamespace = n.oid AND n.nspname = c.table_schema
LEFT JOIN pg_description d ON d.objoid = cl.oid AND d.classoid = 'pg_class'::regclass AND d.objsubid = c.ordinal_position
LEFT JOIN pg_attribute a ON a.attrelid = cl.oid AND a.attname = c.column_name
LEFT JOIN pg_attrdef ad ON ad.adrelid = a.attrelid AND ad.adnum = a.attnum
LEFT JOIN pg_type dt ON dt.oid = a.atttypid
LEFT JOIN pg_namespace dn ON dt.typnamespace = dn.oid
LEFT JOIN pg_type et ON dt.typelem = et.oid
LEFT JOIN pg_namespace en ON et.typnamespace = en.oid
WHERE
c.table_schema NOT IN ('information_schema', 'pg_catalog', 'pg_toast')
AND c.table_schema NOT LIKE 'pg_temp_%'
AND c.table_schema NOT LIKE 'pg_toast_temp_%'
ORDER BY c.table_schema, c.table_name, c.ordinal_position;
cb.table_schema,
cb.table_name,
cb.column_name,
cb.ordinal_position,
-- Use the column_default from LATERAL join which has proper search_path set
ge.column_default,
cb.is_nullable,
cb.data_type,
cb.character_maximum_length,
cb.numeric_precision,
cb.numeric_scale,
cb.udt_name,
cb.column_comment,
cb.resolved_type,
cb.is_identity,
cb.identity_generation,
cb.identity_start,
cb.identity_increment,
cb.identity_maximum,
cb.identity_minimum,
cb.identity_cycle,
cb.attgenerated,
-- Use LATERAL join to guarantee execution order:
-- 1. set_config sets search_path to only the table's schema
Copy link

Copilot AI Dec 30, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The comment states "set_config sets search_path to only the table's schema" but the actual code on line 158 sets the search_path to 'pg_catalog', not the table's schema. This comment should be updated to match the implementation, similar to the comment on lines 259-262.

Suggested change
-- 1. set_config sets search_path to only the table's schema
-- 1. set_config sets search_path to 'pg_catalog' to force schema qualification

Copilot uses AI. Check for mistakes.
-- 2. pg_get_expr then uses that search_path
-- This ensures cross-schema type references in column defaults and generated columns
-- are properly qualified (Issue #218)
ge.generated_expr
FROM column_base cb
LEFT JOIN LATERAL (
SELECT
-- Set search_path to only pg_catalog to force pg_get_expr to include schema qualifiers
-- for all user-defined types and functions. The normalization code will then strip
-- same-schema function qualifiers while preserving type qualifiers (Issue #218)
set_config('search_path', 'pg_catalog', true) as dummy,
CASE
WHEN cb.attgenerated = 's' THEN NULL -- Generated columns don't have defaults
ELSE COALESCE(pg_get_expr(cb.adbin, cb.adrelid), cb.column_default)
END as column_default,
CASE
WHEN cb.attgenerated = 's' THEN pg_get_expr(cb.adbin, cb.adrelid)
ELSE NULL
END as generated_expr
) ge ON true
ORDER BY cb.table_schema, cb.table_name, cb.ordinal_position;

-- GetColumnsForSchema retrieves all columns for tables in a specific schema
-- name: GetColumnsForSchema :many
Expand Down Expand Up @@ -195,10 +237,8 @@ SELECT
cb.table_name,
cb.column_name,
cb.ordinal_position,
CASE
WHEN cb.attgenerated = 's' THEN NULL -- Generated columns don't have defaults
ELSE COALESCE(pg_get_expr(cb.adbin, cb.adrelid), cb.column_default)
END AS column_default,
-- Use the column_default from LATERAL join which has proper search_path set
ge.column_default,
cb.is_nullable,
cb.data_type,
cb.character_maximum_length,
Expand All @@ -216,14 +256,15 @@ SELECT
cb.identity_cycle,
cb.attgenerated,
-- Use LATERAL join to guarantee execution order:
-- 1. set_config sets search_path to only the table's schema
-- 2. pg_get_expr then uses that search_path
-- This ensures cross-schema function references in generated columns are qualified
-- 1. set_config sets search_path to only pg_catalog
-- 2. pg_get_expr then uses that search_path and includes schema qualifiers for user types
-- This ensures type references in column defaults and generated columns are properly
-- qualified (Issue #218). The normalization code strips same-schema function qualifiers.
--
-- NOTE: The 'dummy' column in the LATERAL subquery forces set_config to execute
-- before pg_get_expr. PostgreSQL evaluates SELECT columns left-to-right within
Comment on lines 264 to 265
Copy link

Copilot AI Dec 30, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The comment refers to "'true' in the WHERE clause" but the actual SQL uses "ON true" in the LATERAL join condition (line 291), not a WHERE clause. The comment should be corrected to say "ON condition" instead of "WHERE clause".

Copilot uses AI. Check for mistakes.
-- a single query level. The LATERAL join guarantees this happens row-by-row,
-- and 'true' in the WHERE clause ensures set_config is called for every row.
-- and 'ON true' in the join condition ensures the LATERAL subquery executes for every row.
-- This pattern mirrors GetViewsForSchema (line 959-963) for consistency.
--
-- Alternative considered: Create a custom PostgreSQL function wrapping pg_get_expr
Expand All @@ -235,7 +276,14 @@ SELECT
FROM column_base cb
LEFT JOIN LATERAL (
SELECT
set_config('search_path', cb.table_schema || ', pg_catalog', true) as dummy,
-- Set search_path to only pg_catalog to force pg_get_expr to include schema qualifiers
-- for all user-defined types and functions. The normalization code will then strip
-- same-schema function qualifiers while preserving type qualifiers (Issue #218)
set_config('search_path', 'pg_catalog', true) as dummy,
CASE
WHEN cb.attgenerated = 's' THEN NULL -- Generated columns don't have defaults
ELSE COALESCE(pg_get_expr(cb.adbin, cb.adrelid), cb.column_default)
END as column_default,
CASE
WHEN cb.attgenerated = 's' THEN pg_get_expr(cb.adbin, cb.adrelid)
ELSE NULL
Expand Down
Loading