[bugfix](paimon)fix paimon testcases (#30514)

1. set default timezone
2. not supported `char` type to pushdown

be/src/vec/exec/format/orc/vorc_reader.cpp

@@ -137,7 +137,7 @@ void ORCFileInputStream::read(void* buf, uint64_t length, uint64_t offset) {
 OrcReader::OrcReader(RuntimeProfile* profile, RuntimeState* state,
                      const TFileScanRangeParams& params, const TFileRangeDesc& range,
                      size_t batch_size, const std::string& ctz, io::IOContext* io_ctx,
-                     bool enable_lazy_mat)
+                     bool enable_lazy_mat, std::vector<orc::TypeKind>* unsupported_pushdown_types)
         : _profile(profile),
           _state(state),
           _scan_params(params),
@@ -148,7 +148,8 @@ OrcReader::OrcReader(RuntimeProfile* profile, RuntimeState* state,
           _ctz(ctz),
           _io_ctx(io_ctx),
           _enable_lazy_mat(enable_lazy_mat),
-          _is_dict_cols_converted(false) {
+          _is_dict_cols_converted(false),
+          _unsupported_pushdown_types(unsupported_pushdown_types) {
     TimezoneUtils::find_cctz_time_zone(ctz, _time_zone);
     VecDateTimeValue t;
     t.from_unixtime(0, ctz);
@@ -524,8 +525,20 @@ std::tuple<bool, orc::Literal> convert_to_orc_literal(const orc::Type* type, con
 
 template <PrimitiveType primitive_type>
 std::vector<OrcPredicate> value_range_to_predicate(
-        const ColumnValueRange<primitive_type>& col_val_range, const orc::Type* type) {
+        const ColumnValueRange<primitive_type>& col_val_range, const orc::Type* type,
+        std::vector<orc::TypeKind>* unsupported_pushdown_types) {
     std::vector<OrcPredicate> predicates;
+
+    if (unsupported_pushdown_types != nullptr) {
+        for (vector<orc::TypeKind>::iterator it = unsupported_pushdown_types->begin();
+             it != unsupported_pushdown_types->end(); ++it) {
+            if (*it == type->getKind()) {
+                // Unsupported type
+                return predicates;
+            }
+        }
+    }
+
     orc::PredicateDataType predicate_data_type;
     auto type_it = TYPEKIND_TO_PREDICATE_TYPE.find(type->getKind());
     if (type_it == TYPEKIND_TO_PREDICATE_TYPE.end()) {
@@ -667,8 +680,8 @@ bool OrcReader::_init_search_argument(
         }
         std::visit(
                 [&](auto& range) {
-                    std::vector<OrcPredicate> value_predicates =
-                            value_range_to_predicate(range, type_it->second);
+                    std::vector<OrcPredicate> value_predicates = value_range_to_predicate(
+                            range, type_it->second, _unsupported_pushdown_types);
                     for (auto& range_predicate : value_predicates) {
                         predicates.emplace_back(range_predicate);
                     }

be/src/vec/exec/format/orc/vorc_reader.h

@@ -130,7 +130,8 @@ class OrcReader : public GenericReader {
 
     OrcReader(RuntimeProfile* profile, RuntimeState* state, const TFileScanRangeParams& params,
               const TFileRangeDesc& range, size_t batch_size, const std::string& ctz,
-              io::IOContext* io_ctx, bool enable_lazy_mat = true);
+              io::IOContext* io_ctx, bool enable_lazy_mat = true,
+              std::vector<orc::TypeKind>* unsupported_pushdown_types = nullptr);
 
     OrcReader(const TFileScanRangeParams& params, const TFileRangeDesc& range,
               const std::string& ctz, io::IOContext* io_ctx, bool enable_lazy_mat = true);
@@ -571,6 +572,7 @@ class OrcReader : public GenericReader {
     std::unique_ptr<orc::StringDictFilter> _string_dict_filter;
     bool _is_dict_cols_converted;
     bool _has_complex_type = false;
+    std::vector<orc::TypeKind>* _unsupported_pushdown_types;
 };
 
 class ORCFileInputStream : public orc::InputStream {

be/src/vec/exec/scan/vfile_scanner.cpp

@@ -785,9 +785,20 @@ Status VFileScanner::_get_next_reader() {
             break;
         }
         case TFileFormatType::FORMAT_PARQUET: {
+            static const cctz::time_zone utc0 = cctz::utc_time_zone();
+            cctz::time_zone* tz;
+            if (range.__isset.table_format_params &&
+                range.table_format_params.table_format_type == "paimon") {
+                // The timestmap generated by paimon does not carry metadata information (e.g., isAdjustToUTC, etc.),
+                // and the stored data is UTC0 by default, so it is directly set to the UTC time zone.
+                // In version 0.7, paimon fixed this issue and can remove the judgment here
+                tz = const_cast<cctz::time_zone*>(&utc0);
+            } else {
+                tz = const_cast<cctz::time_zone*>(&_state->timezone_obj());
+            }
             std::unique_ptr<ParquetReader> parquet_reader = ParquetReader::create_unique(
-                    _profile, *_params, range, _state->query_options().batch_size,
-                    const_cast<cctz::time_zone*>(&_state->timezone_obj()), _io_ctx.get(), _state,
+                    _profile, *_params, range, _state->query_options().batch_size, tz,
+                    _io_ctx.get(), _state,
                     config::max_external_file_meta_cache_num <= 0
                             ? nullptr
                             : ExecEnv::GetInstance()->file_meta_cache(),
@@ -825,9 +836,17 @@ Status VFileScanner::_get_next_reader() {
             break;
         }
         case TFileFormatType::FORMAT_ORC: {
+            std::vector<orc::TypeKind>* unsupported_pushdown_types = nullptr;
+            if (range.__isset.table_format_params &&
+                range.table_format_params.table_format_type == "paimon") {
+                static std::vector<orc::TypeKind> paimon_unsupport_type =
+                        std::vector<orc::TypeKind> {orc::TypeKind::CHAR};
+                unsupported_pushdown_types = &paimon_unsupport_type;
+            }
             std::unique_ptr<OrcReader> orc_reader = OrcReader::create_unique(
                     _profile, _state, *_params, range, _state->query_options().batch_size,
-                    _state->timezone(), _io_ctx.get(), _state->query_options().enable_orc_lazy_mat);
+                    _state->timezone(), _io_ctx.get(), _state->query_options().enable_orc_lazy_mat,
+                    unsupported_pushdown_types);
             if (push_down_predicates) {
                 RETURN_IF_ERROR(_process_late_arrival_conjuncts());
             }

fe/fe-core/src/main/java/org/apache/doris/planner/external/paimon/PaimonValueConverter.java

@@ -23,9 +23,7 @@
 import org.apache.doris.analysis.FloatLiteral;
 import org.apache.doris.analysis.IntLiteral;
 import org.apache.doris.analysis.LiteralExpr;
-import org.apache.doris.analysis.StringLiteral;
 
-import com.google.common.base.Strings;
 import org.apache.paimon.data.BinaryString;
 import org.apache.paimon.data.Decimal;
 import org.apache.paimon.data.Timestamp;
@@ -64,10 +62,8 @@ public BinaryString visit(VarCharType varCharType) {
     }
 
     public BinaryString visit(CharType charType) {
-        if (expr instanceof StringLiteral) {
-            StringLiteral stringLiteral = (StringLiteral) expr;
-            return BinaryString.fromString(Strings.padEnd(stringLiteral.getStringValue(), charType.getLength(), ' '));
-        }
+        // Currently, Paimon does not support predicate push-down for char
+        // ref: org.apache.paimon.predicate.PredicateBuilder.convertJavaObject
         return null;
     }