feat: Google BigQuery CAST with FORMAT clause

Signed-off-by: Andreas Reichel <[email protected]>
JSQLParser · Mar 31, 2024 · 0d813f0 · 0d813f0
1 parent 236793a
commit 0d813f0
Show file tree

Hide file tree

Showing 5 changed files with 201 additions and 2 deletions.
diff --git a/src/main/java/net/sf/jsqlparser/expression/CastExpression.java b/src/main/java/net/sf/jsqlparser/expression/CastExpression.java
@@ -23,6 +23,10 @@ public class CastExpression extends ASTNodeAccessImpl implements Expression {
     private ColDataType colDataType = null;
     private ArrayList<ColumnDefinition> columnDefinitions = new ArrayList<>();
 
+    // BigQuery specific FORMAT clause:
+    // https://cloud.google.com/bigquery/docs/reference/standard-sql/conversion_functions#cast_as_date
+    private String format = null;
+
     public CastExpression(String keyword, Expression leftExpression, String dataType) {
         this.keyword = keyword;
         this.leftExpression = leftExpression;
@@ -89,13 +93,26 @@ public void setUseCastKeyword(boolean useCastKeyword) {
         }
     }
 
+    public String getFormat() {
+        return format;
+    }
+
+    public CastExpression setFormat(String format) {
+        this.format = format;
+        return this;
+    }
+
     @Override
     public String toString() {
+        String formatStr = format != null && !format.isEmpty()
+                ? " FORMAT " + format
+                : "";
         if (keyword != null && !keyword.isEmpty()) {
             return columnDefinitions.size() > 1
                     ? keyword + "(" + leftExpression + " AS ROW("
-                            + Select.getStringList(columnDefinitions) + "))"
-                    : keyword + "(" + leftExpression + " AS " + colDataType.toString() + ")";
+                            + Select.getStringList(columnDefinitions) + ")" + formatStr + ")"
+                    : keyword + "(" + leftExpression + " AS " + colDataType.toString() + formatStr
+                            + ")";
         } else {
             return leftExpression + "::" + colDataType.toString();
         }

diff --git a/src/main/java/net/sf/jsqlparser/util/deparser/ExpressionDeParser.java b/src/main/java/net/sf/jsqlparser/util/deparser/ExpressionDeParser.java
@@ -704,13 +704,18 @@ public void visit(BitwiseXor bitwiseXor) {
     @Override
     public void visit(CastExpression cast) {
         if (cast.isUseCastKeyword()) {
+            String formatStr = cast.getFormat() != null && !cast.getFormat().isEmpty()
+                    ? " FORMAT " + cast.getFormat()
+                    : "";
+
             buffer.append(cast.keyword).append("(");
             cast.getLeftExpression().accept(this);
             buffer.append(" AS ");
             buffer.append(
                     cast.getColumnDefinitions().size() > 1
                             ? "ROW(" + Select.getStringList(cast.getColumnDefinitions()) + ")"
                             : cast.getColDataType().toString());
+            buffer.append(formatStr);
             buffer.append(")");
         } else {
             cast.getLeftExpression().accept(this);

diff --git a/src/main/jjtree/net/sf/jsqlparser/parser/JSqlParserCC.jjt b/src/main/jjtree/net/sf/jsqlparser/parser/JSqlParserCC.jjt
@@ -5033,6 +5033,7 @@ CastExpression CastExpression():
     ColDataType type;
     Expression expression = null;
     boolean    useCastKeyword;
+    Token formatCharLiteral;
 }
 {
     (
@@ -5061,6 +5062,11 @@ CastExpression CastExpression():
         |
         type=ColDataType() { retval.setColDataType(type); }
     )
+
+    // BigQuery FORMAT clause
+    // https://cloud.google.com/bigquery/docs/reference/standard-sql/conversion_functions#cast_as_date
+    [ <K_FORMAT> formatCharLiteral=<S_CHAR_LITERAL> { retval.setFormat(formatCharLiteral.image); } ]
+
     ")"
 
     {

diff --git a/src/test/java/net/sf/jsqlparser/expression/CaseExpressionTest.java b/src/test/java/net/sf/jsqlparser/expression/CaseExpressionTest.java
@@ -160,4 +160,10 @@ void testPerformanceIssue1889() throws JSQLParserException {
 
         TestUtils.assertSqlCanBeParsedAndDeparsed(sqlStr, true);
     }
+
+    @Test
+    void testFormatClause() throws JSQLParserException {
+        String sqlStr = "SELECT CAST('18-12-03' AS DATE FORMAT 'YY-MM-DD') AS string_to_date";
+        TestUtils.assertSqlCanBeParsedAndDeparsed(sqlStr, true);
+    }
 }
diff --git a/src/test/java/net/sf/jsqlparser/statement/select/NestedBracketsPerformanceTest.java b/src/test/java/net/sf/jsqlparser/statement/select/NestedBracketsPerformanceTest.java
@@ -10,6 +10,7 @@
 package net.sf.jsqlparser.statement.select;
 
 import net.sf.jsqlparser.JSQLParserException;
+import net.sf.jsqlparser.parser.CCJSqlParserUtil;
 import org.junit.jupiter.api.Disabled;
 import org.junit.jupiter.api.Test;
 import org.junit.jupiter.api.Timeout;
@@ -353,4 +354,168 @@ public void testDeepFunctionParameters() throws JSQLParserException {
 
         assertSqlCanBeParsedAndDeparsed(sqlStr, true);
     }
+
+    @Test
+    @Disabled
+    void testIssue1983() throws JSQLParserException {
+        String sqlStr = "INSERT INTO\n" +
+                "C01_INDIV_TELBK_CUST_INFO_H_T2 (PARTY_ID, PARTY_SIGN_STAT_CD, SIGN_TM, CLOSE_TM)\n"
+                +
+                "SELECT\n" +
+                "A1.PARTY_ID,\n" +
+                "A1.PARTY_SIGN_STAT_CD,\n" +
+                "CAST(\n" +
+                "(\n" +
+                "CASE\n" +
+                "WHEN A1.SIGN_TM IS NULL\n" +
+                "OR A1.SIGN_TM = '' THEN CAST(\n" +
+                "CAST(\n" +
+                "CAST('ATkkIVQJZm' AS DATE FORMAT 'YYYYMMDD') AS DATE\n" +
+                ") || ' 00:00:00' AS TIMESTAMP\n" +
+                ")\n" +
+                "WHEN CHARACTERS (TRIM(A1.SIGN_TM)) <> 19\n" +
+                "OR SUBSTR (TRIM(A1.SIGN_TM), 1, 1) < '0'\n" +
+                "OR SUBSTR (TRIM(A1.SIGN_TM), 1, 1) > '9'\n" +
+                "OR SUBSTR (TRIM(A1.SIGN_TM), 2, 1) < '0'\n" +
+                "OR SUBSTR (TRIM(A1.SIGN_TM), 2, 1) > '9'\n" +
+                "OR SUBSTR (TRIM(A1.SIGN_TM), 3, 1) < '0'\n" +
+                "OR SUBSTR (TRIM(A1.SIGN_TM), 3, 1) > '9'\n" +
+                "OR SUBSTR (TRIM(A1.SIGN_TM), 4, 1) < '0'\n" +
+                "OR SUBSTR (TRIM(A1.SIGN_TM), 4, 1) > '9'\n" +
+                "OR SUBSTR (TRIM(A1.SIGN_TM), 6, 1) < '0'\n" +
+                "OR SUBSTR (TRIM(A1.SIGN_TM), 6, 1) > '1'\n" +
+                "OR SUBSTR (TRIM(A1.SIGN_TM), 7, 1) < '0'\n" +
+                "OR SUBSTR (TRIM(A1.SIGN_TM), 7, 1) > '9'\n" +
+                "OR SUBSTR (TRIM(A1.SIGN_TM), 9, 1) < '0'\n" +
+                "OR SUBSTR (TRIM(A1.SIGN_TM), 9, 1) > '3'\n" +
+                "OR SUBSTR (TRIM(A1.SIGN_TM), 10, 1) < '0'\n" +
+                "OR SUBSTR (TRIM(A1.SIGN_TM), 10, 1) > '9'\n" +
+                "OR SUBSTR (TRIM(A1.SIGN_TM), 1, 4) = '0000'\n" +
+                "OR SUBSTR (TRIM(A1.SIGN_TM), 6, 2) = '00'\n" +
+                "OR SUBSTR (TRIM(A1.SIGN_TM), 9, 2) = '00'\n" +
+                "OR SUBSTR (TRIM(A1.SIGN_TM), 1, 1) = '0' THEN CAST(\n" +
+                "CAST(\n" +
+                "CAST('cDXtwdFyky' AS DATE FORMAT 'YYYYMMDD') AS DATE\n" +
+                ") || ' 00:00:00' AS TIMESTAMP\n" +
+                ")\n" +
+                "ELSE (\n" +
+                "CASE\n" +
+                "WHEN (\n" +
+                "CAST(SUBSTR (TRIM(A1.SIGN_TM), 9, 2) AS INTEGER) < 29\n" +
+                "AND SUBSTR (TRIM(A1.SIGN_TM), 6, 2) = '02'\n" +
+                ")\n" +
+                "OR (\n" +
+                "CAST(SUBSTR (TRIM(A1.SIGN_TM), 9, 2) AS INTEGER) < 31\n" +
+                "AND SUBSTR (TRIM(A1.SIGN_TM), 6, 2) <> '02'\n" +
+                "AND SUBSTR (TRIM(A1.SIGN_TM), 6, 2) <= 12\n" +
+                ")\n" +
+                "OR (\n" +
+                "CAST(SUBSTR (TRIM(A1.SIGN_TM), 9, 2) AS INTEGER) = 31\n" +
+                "AND SUBSTR (TRIM(A1.SIGN_TM), 6, 2) IN ('01', '03', '05', '07', '08', '10', '12')\n"
+                +
+                ") THEN CAST(A1.SIGN_TM AS TIMESTAMP)\n" +
+                "WHEN SUBSTR (TRIM(A1.SIGN_TM), 6, 2) || SUBSTR (TRIM(A1.SIGN_TM), 9, 2) = '0229'\n"
+                +
+                "AND (\n" +
+                "CAST(SUBSTR (TRIM(A1.SIGN_TM), 1, 4) AS INTEGER) MOD 400 = 0\n" +
+                "OR (\n" +
+                "CAST(SUBSTR (TRIM(A1.SIGN_TM), 1, 4) AS INTEGER) MOD 4 = 0\n" +
+                "AND CAST(SUBSTR (TRIM(A1.SIGN_TM), 1, 4) AS INTEGER) MOD 100 <> 0\n" +
+                ")\n" +
+                ") THEN CAST(A1.SIGN_TM AS TIMESTAMP)\n" +
+                "ELSE CAST(\n" +
+                "CAST(\n" +
+                "CAST('cDXtwdFyky' AS DATE FORMAT 'YYYYMMDD') AS DATE\n" +
+                ") || ' 00:00:00' AS TIMESTAMP\n" +
+                ")\n" +
+                "END\n" +
+                ")\n" +
+                "END\n" +
+                ") AS DATE FORMAT 'YYYYMMDD'\n" +
+                "),\n" +
+                "CAST(\n" +
+                "(\n" +
+                "CASE\n" +
+                "WHEN A1.CLOSE_TM IS NULL\n" +
+                "OR A1.CLOSE_TM = '' THEN CAST(\n" +
+                "CAST(\n" +
+                "CAST('ATkkIVQJZm' AS DATE FORMAT 'YYYYMMDD') AS DATE\n" +
+                ") || ' 00:00:00' AS TIMESTAMP\n" +
+                ")\n" +
+                "WHEN CHARACTERS (TRIM(A1.CLOSE_TM)) <> 19\n" +
+                "OR SUBSTR (TRIM(A1.CLOSE_TM), 1, 1) < '0'\n" +
+                "OR SUBSTR (TRIM(A1.CLOSE_TM), 1, 1) > '9'\n" +
+                "OR SUBSTR (TRIM(A1.CLOSE_TM), 2, 1) < '0'\n" +
+                "OR SUBSTR (TRIM(A1.CLOSE_TM), 2, 1) > '9'\n" +
+                "OR SUBSTR (TRIM(A1.CLOSE_TM), 3, 1) < '0'\n" +
+                "OR SUBSTR (TRIM(A1.CLOSE_TM), 3, 1) > '9'\n" +
+                "OR SUBSTR (TRIM(A1.CLOSE_TM), 4, 1) < '0'\n" +
+                "OR SUBSTR (TRIM(A1.CLOSE_TM), 4, 1) > '9'\n" +
+                "OR SUBSTR (TRIM(A1.CLOSE_TM), 6, 1) < '0'\n" +
+                "OR SUBSTR (TRIM(A1.CLOSE_TM), 6, 1) > '1'\n" +
+                "OR SUBSTR (TRIM(A1.CLOSE_TM), 7, 1) < '0'\n" +
+                "OR SUBSTR (TRIM(A1.CLOSE_TM), 7, 1) > '9'\n" +
+                "OR SUBSTR (TRIM(A1.CLOSE_TM), 9, 1) < '0'\n" +
+                "OR SUBSTR (TRIM(A1.CLOSE_TM), 9, 1) > '3'\n" +
+                "OR SUBSTR (TRIM(A1.CLOSE_TM), 10, 1) < '0'\n" +
+                "OR SUBSTR (TRIM(A1.CLOSE_TM), 10, 1) > '9'\n" +
+                "OR SUBSTR (TRIM(A1.CLOSE_TM), 1, 4) = '0000'\n" +
+                "OR SUBSTR (TRIM(A1.CLOSE_TM), 6, 2) = '00'\n" +
+                "OR SUBSTR (TRIM(A1.CLOSE_TM), 9, 2) = '00'\n" +
+                "OR SUBSTR (TRIM(A1.CLOSE_TM), 1, 1) = '0' THEN CAST(\n" +
+                "CAST(\n" +
+                "CAST('cDXtwdFyky' AS DATE FORMAT 'YYYYMMDD') AS DATE\n" +
+                ") || ' 00:00:00' AS TIMESTAMP\n" +
+                ")\n" +
+                "ELSE (\n" +
+                "CASE\n" +
+                "WHEN (\n" +
+                "CAST(SUBSTR (TRIM(A1.CLOSE_TM), 9, 2) AS INTEGER) < 29\n" +
+                "AND SUBSTR (TRIM(A1.CLOSE_TM), 6, 2) = '02'\n" +
+                ")\n" +
+                "OR (\n" +
+                "CAST(SUBSTR (TRIM(A1.CLOSE_TM), 9, 2) AS INTEGER) < 31\n" +
+                "AND SUBSTR (TRIM(A1.CLOSE_TM), 6, 2) <> '02'\n" +
+                "AND SUBSTR (TRIM(A1.CLOSE_TM), 6, 2) <= 12\n" +
+                ")\n" +
+                "OR (\n" +
+                "CAST(SUBSTR (TRIM(A1.CLOSE_TM), 9, 2) AS INTEGER) = 31\n" +
+                "AND SUBSTR (TRIM(A1.CLOSE_TM), 6, 2) IN ('01', '03', '05', '07', '08', '10', '12')\n"
+                +
+                ") THEN CAST(A1.CLOSE_TM AS TIMESTAMP)\n" +
+                "WHEN SUBSTR (TRIM(A1.CLOSE_TM), 6, 2) || SUBSTR (TRIM(A1.CLOSE_TM), 9, 2) = '0229'\n"
+                +
+                "AND (\n" +
+                "CAST(SUBSTR (TRIM(A1.CLOSE_TM), 1, 4) AS INTEGER) MOD 400 = 0\n" +
+                "OR (\n" +
+                "CAST(SUBSTR (TRIM(A1.CLOSE_TM), 1, 4) AS INTEGER) MOD 4 = 0\n" +
+                "AND CAST(SUBSTR (TRIM(A1.CLOSE_TM), 1, 4) AS INTEGER) MOD 100 <> 0\n" +
+                ")\n" +
+                ") THEN CAST(A1.CLOSE_TM AS TIMESTAMP)\n" +
+                "ELSE CAST(\n" +
+                "CAST(\n" +
+                "CAST('cDXtwdFyky' AS DATE FORMAT 'YYYYMMDD') AS DATE\n" +
+                ") || ' 00:00:00' AS TIMESTAMP\n" +
+                ")\n" +
+                "END\n" +
+                ")\n" +
+                "END\n" +
+                ") AS DATE FORMAT 'YYYYMMDD'\n" +
+                ")\n" +
+                "FROM\n" +
+                "T01_PTY_SIGN_H_T1 A1\n" +
+                "WHERE\n" +
+                "A1.PARTY_SIGN_TYPE_CD = 'CD_021'\n" +
+                "AND A1.ST_DT <= CAST('LDBCGtCIyo' AS DATE FORMAT 'YYYYMMDD')\n" +
+                "AND A1.END_DT > CAST('LDBCGtCIyo' AS DATE FORMAT 'YYYYMMDD')\n" +
+                "GROUP BY\n" +
+                "1,\n" +
+                "2,\n" +
+                "3,\n" +
+                "4";
+        CCJSqlParserUtil.parse(sqlStr, parser -> parser
+                .withSquareBracketQuotation(false)
+                .withAllowComplexParsing(true)
+                .withTimeOut(60000));
+    }
 }