pkg/format/body.go

@@ -441,13 +441,9 @@ func formatBodyStatements(text string, st config.Style) string {
 			stmtLines = joinThenToCondition(stmt)
 		}
 
-		for i, ll := range stmtLines {
-			if i == 0 || ll.indent == "" {
-				result.WriteString(baseIndent)
-			} else {
-				result.WriteString(ll.indent)
-			}
-			result.WriteString(ll.text)
+		formattedLines := formatBodyStmtLines(stmtLines, baseIndent, st)
+		for _, line := range formattedLines {
+			result.WriteString(line)
 			result.WriteString(nl)
 		}
 
@@ -574,6 +570,131 @@ func formatBodyStatements(text string, st config.Style) string {
 	return result.String()
 }
 
+// formatBodyStmtLines formats one flushed PL/pgSQL statement at its contextual
+// base indent. Multi-line UPDATE/DELETE statements inside PL/pgSQL get their
+// top-level SET/WHERE/AND/OR clauses realigned under the statement while nested
+// subqueries keep their original indentation. Non-DML statements keep
+// continuation indentation, except that standalone structural keywords such as
+// THEN are aligned with the block opener.
+func formatBodyStmtLines(lines []bline, baseIndent string, st config.Style) []string {
+	if len(lines) == 0 {
+		return nil
+	}
+
+	if looksLikeMultiLineBodyDML(lines) {
+		return reindentBodyDML(lines, baseIndent, st)
+	}
+
+	out := make([]string, 0, len(lines))
+	for i, ll := range lines {
+		text := ll.text
+		indent := baseIndent
+		if i > 0 && ll.indent != "" && !isStandaloneBodyKeyword(ll.text, "then", "else", "elsif", "elseif") {
+			indent = ll.indent
+		}
+		out = append(out, indent+text)
+	}
+	return out
+}
+
+func looksLikeMultiLineBodyDML(lines []bline) bool {
+	if len(lines) < 2 {
+		return false
+	}
+	kw := lowerASCII(firstBodyKeyword(lines[0].text))
+	return kw == "update" || kw == "delete"
+}
+
+func reindentBodyDML(lines []bline, baseIndent string, st config.Style) []string {
+	out := make([]string, 0, len(lines)+1)
+	afterWhere := false
+	parenDepth := 0
+	for i, ll := range lines {
+		text := strings.TrimRight(ll.text, " 	")
+		lineDepth := parenDepth
+		kw := lowerASCII(firstBodyKeyword(text))
+		if afterWhere && lineDepth == 0 && kw != "and" && kw != "or" {
+			out = append(out, baseIndent+st.Indent+st.Indent+strings.TrimSpace(text))
+			afterWhere = false
+			updateBodyParenDepth(text, &parenDepth)
+			continue
+		}
+		if lineDepth == 0 && (kw == "set" || kw == "where" || kw == "values" || kw == "returning") {
+			if kw == "where" {
+				whereText := strings.TrimSpace(text)
+				fields := strings.Fields(whereText)
+				nextKw := ""
+				if i+1 < len(lines) {
+					nextKw = lowerASCII(firstBodyKeyword(lines[i+1].text))
+				}
+				if len(fields) > 1 && (nextKw == "and" || nextKw == "or") {
+					out = append(out, baseIndent+fields[0])
+					out = append(out, baseIndent+st.Indent+st.Indent+strings.TrimSpace(whereText[len(fields[0]):]))
+					afterWhere = false
+					continue
+				}
+				afterWhere = len(fields) == 1
+			}
+			out = append(out, baseIndent+strings.TrimSpace(text))
+			continue
+		}
+		if lineDepth == 0 && (kw == "and" || kw == "or") {
+			out = append(out, baseIndent+st.Indent+strings.TrimSpace(text))
+			afterWhere = false
+			updateBodyParenDepth(text, &parenDepth)
+			continue
+		}
+		if i == 0 {
+			out = append(out, baseIndent+strings.TrimSpace(text))
+		} else if ll.indent != "" {
+			out = append(out, ll.indent+strings.TrimSpace(text))
+		} else {
+			out = append(out, baseIndent+strings.TrimSpace(text))
+		}
+		afterWhere = false
+		updateBodyParenDepth(text, &parenDepth)
+	}
+	return out
+}
+
+func updateBodyParenDepth(s string, depth *int) {
+	for _, tok := range lexer.Lex(s) {
+		switch tok.Kind {
+		case lexer.LParen, lexer.LBracket:
+			(*depth)++
+		case lexer.RParen, lexer.RBracket:
+			if *depth > 0 {
+				(*depth)--
+			}
+		}
+	}
+}
+
+func significantBodyTokens(s string) []cst.Tok {
+	var toks []cst.Tok
+	for _, tok := range lexer.Lex(s) {
+		if tok.IsTrivia() || tok.Kind == lexer.EOF {
+			continue
+		}
+		toks = append(toks, cst.Tok{Tok: tok})
+	}
+	return toks
+}
+
+func isStandaloneBodyKeyword(s string, kws ...string) bool {
+	toks := significantBodyTokens(s)
+	if len(toks) != 1 || toks[0].Tok.Kind != lexer.Ident {
+		return false
+	}
+	low := lowerASCII(toks[0].Tok.Text)
+	for _, kw := range kws {
+		if low == kw {
+			return true
+		}
+	}
+	return false
+}
+
 // joinThenToCondition merges a THEN line (on its own bline) into the preceding
 // condition line when plpgsql_if_then_newline is false.
 func joinThenToCondition(lines []bline) []bline {

pkg/format/format_test.go

@@ -107,6 +107,66 @@ func TestFormatMmProcBroken(t *testing.T) {
 	}
 }
 
+func TestFormatIssue1PLpgSQLIndenting(t *testing.T) {
+	src := "CREATE FUNCTION f() RETURNS void LANGUAGE plpgsql AS $$\n" +
+		"DECLARE\n" +
+		"  r_lp record;\n" +
+		"BEGIN\n" +
+		"   if r_lp.total > 0\n" +
+		"      and r_lp.totaldone >= r_lp.total\n" +
+		"    then\n" +
+		"    update core.process u\n" +
+		"      set status = 'done'\n" +
+		"      where u.rid_process = r_lp.rid_process\n" +
+		"        and nv(u.status) <> 'done';\n" +
+		"  elsif r_lp.total > 0\n" +
+		"    then\n" +
+		"    update core.process u\n" +
+		"      set status = 'open'\n" +
+		"      where u.rid_process = r_lp.rid_process\n" +
+		"        and nv(u.status) <> 'open';\n" +
+		"\n" +
+		"  end if;\n" +
+		"$$;\n"
+
+	want := "CREATE FUNCTION f(\n" +
+		")\n" +
+		"RETURNS void\n" +
+		"LANGUAGE plpgsql\n" +
+		"AS\n" +
+		"$$\n" +
+		"DECLARE\n" +
+		"  r_lp record;\n" +
+		"BEGIN\n" +
+		"  if r_lp.total > 0\n" +
+		"      and r_lp.totaldone >= r_lp.total\n" +
+		"  then\n" +
+		"    update core.process u\n" +
+		"    set status = 'done'\n" +
+		"    where\n" +
+		"        u.rid_process = r_lp.rid_process\n" +
+		"      and nv(u.status) <> 'done';\n" +
+		"  elsif r_lp.total > 0\n" +
+		"  then\n" +
+		"    update core.process u\n" +
+		"    set status = 'open'\n" +
+		"    where\n" +
+		"        u.rid_process = r_lp.rid_process\n" +
+		"      and nv(u.status) <> 'open';\n" +
+		"\n" +
+		"  end if;\n" +
+		"$$;\n"
+
+	got := format(src)
+	if got != want {
+		t.Errorf("issue #1 PL/pgSQL indenting\n--- got ---\n%s\n--- want ---\n%s", got, want)
+	}
+	checkDML(t, "issue #1 PL/pgSQL indenting", got)
+	if !semanticallyEqual(src, got) {
+		t.Errorf("issue #1 PL/pgSQL indenting changed semantics")
+	}
+}
+
 func TestCorpusIdempotentAndSafe(t *testing.T) {
 	dir := filepath.Join("..", "..", "testdata", "corpus")
 	entries, err := os.ReadDir(dir)