This is an automated email from the ASF dual-hosted git repository. yiguolei pushed a commit to branch branch-2.1 in repository https://gitbox.apache.org/repos/asf/doris.git
commit 203daba19d2fa274dc383e811e348d196c1efb02 Author: xueweizhang <[email protected]> AuthorDate: Thu Feb 1 18:47:30 2024 +0800 [fix](outfile) fix outfile csv did not write json column with string (#29067) --- .../vec/data_types/serde/data_type_jsonb_serde.cpp | 6 +++- .../data/export_p0/test_outfile_csv_with_names.out | 40 +++++++++++----------- .../export_p0/test_outfile_csv_with_names.groovy | 17 +++++---- 3 files changed, 35 insertions(+), 28 deletions(-) diff --git a/be/src/vec/data_types/serde/data_type_jsonb_serde.cpp b/be/src/vec/data_types/serde/data_type_jsonb_serde.cpp index 8a6b26e9bd8..2015ed082b1 100644 --- a/be/src/vec/data_types/serde/data_type_jsonb_serde.cpp +++ b/be/src/vec/data_types/serde/data_type_jsonb_serde.cpp @@ -83,7 +83,11 @@ Status DataTypeJsonbSerDe::serialize_one_cell_to_json(const IColumn& column, int const StringRef& s = assert_cast<const ColumnString&>(*ptr).get_data_at(row_num); if (s.size > 0) { - bw.write(s.data, s.size); + std::string str = JsonbToJson::jsonb_to_json_string(s.data, s.size); + bw.write(str.c_str(), str.size()); + } else { + bw.write(NULL_IN_CSV_FOR_ORDINARY_TYPE.c_str(), + strlen(NULL_IN_CSV_FOR_ORDINARY_TYPE.c_str())); } return Status::OK(); } diff --git a/regression-test/data/export_p0/test_outfile_csv_with_names.out b/regression-test/data/export_p0/test_outfile_csv_with_names.out index 63fa1d523c8..5a6badd50e2 100644 --- a/regression-test/data/export_p0/test_outfile_csv_with_names.out +++ b/regression-test/data/export_p0/test_outfile_csv_with_names.out @@ -1,25 +1,25 @@ -- This file is automatically generated. You should know what you did if you want to edit this -- !select_default1 -- -1 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 1 1 true 1 1 1 1.1 1.1 char1 1 -10 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 \N \N \N \N \N \N \N \N \N \N \N -2 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 2 2 true 2 2 2 2.2 2.2 char2 2 -3 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 3 3 true 3 3 3 3.3 3.3 char3 3 -4 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 4 4 true 4 4 4 4.4 4.4 char4 4 -5 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 5 5 true 5 5 5 5.5 5.5 char5 5 -6 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 6 6 true 6 6 6 6.6 6.6 char6 6 -7 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 7 7 true 7 7 7 7.7 7.7 char7 7 -8 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 8 8 true 8 8 8 8.8 8.8 char8 8 -9 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 9 9 true 9 9 9 9.9 9.9 char9 9 +1 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 1 1 true 1 1 1 1.1 1.1 char1 1 {"a":1,"b":"str1"} +10 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 \N \N \N \N \N \N \N \N \N \N \N \N +2 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 2 2 true 2 2 2 2.2 2.2 char2 2 {"a":2,"b":"str2"} +3 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 3 3 true 3 3 3 3.3 3.3 char3 3 {"a":3,"b":"str3"} +4 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 4 4 true 4 4 4 4.4 4.4 char4 4 {"a":4,"b":"str4"} +5 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 5 5 true 5 5 5 5.5 5.5 char5 5 {"a":5,"b":"str5"} +6 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 6 6 true 6 6 6 6.6 6.6 char6 6 {"a":6,"b":"str6"} +7 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 7 7 true 7 7 7 7.7 7.7 char7 7 {"a":7,"b":"str7"} +8 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 8 8 true 8 8 8 8.8 8.8 char8 8 {"a":8,"b":"str8"} +9 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 9 9 true 9 9 9 9.9 9.9 char9 9 {"a":9,"b":"str9"} -- !select_default2 -- -1 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 1 1 true 1 1 1 1.1 1.1 char1 1 -10 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 \N \N \N \N \N \N \N \N \N \N \N -2 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 2 2 true 2 2 2 2.2 2.2 char2 2 -3 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 3 3 true 3 3 3 3.3 3.3 char3 3 -4 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 4 4 true 4 4 4 4.4 4.4 char4 4 -5 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 5 5 true 5 5 5 5.5 5.5 char5 5 -6 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 6 6 true 6 6 6 6.6 6.6 char6 6 -7 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 7 7 true 7 7 7 7.7 7.7 char7 7 -8 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 8 8 true 8 8 8 8.8 8.8 char8 8 -9 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 9 9 true 9 9 9 9.9 9.9 char9 9 +1 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 1 1 true 1 1 1 1.1 1.1 char1 1 {"a":1,"b":"str1"} +10 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 \N \N \N \N \N \N \N \N \N \N \N \N +2 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 2 2 true 2 2 2 2.2 2.2 char2 2 {"a":2,"b":"str2"} +3 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 3 3 true 3 3 3 3.3 3.3 char3 3 {"a":3,"b":"str3"} +4 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 4 4 true 4 4 4 4.4 4.4 char4 4 {"a":4,"b":"str4"} +5 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 5 5 true 5 5 5 5.5 5.5 char5 5 {"a":5,"b":"str5"} +6 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 6 6 true 6 6 6 6.6 6.6 char6 6 {"a":6,"b":"str6"} +7 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 7 7 true 7 7 7 7.7 7.7 char7 7 {"a":7,"b":"str7"} +8 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 8 8 true 8 8 8 8.8 8.8 char8 8 {"a":8,"b":"str8"} +9 2017-10-01 2017-10-01T00:00 2017-10-01 2017-10-01T00:00 2017-10-01T00:00:00.111 2017-10-01T00:00:00.111111 Beijing 9 9 true 9 9 9 9.9 9.9 char9 9 {"a":9,"b":"str9"} diff --git a/regression-test/suites/export_p0/test_outfile_csv_with_names.groovy b/regression-test/suites/export_p0/test_outfile_csv_with_names.groovy index 9c62a994c47..97722100191 100644 --- a/regression-test/suites/export_p0/test_outfile_csv_with_names.groovy +++ b/regression-test/suites/export_p0/test_outfile_csv_with_names.groovy @@ -77,7 +77,8 @@ suite("test_outfile_csv_with_names") { `float_col` float COMMENT "", `double_col` double COMMENT "", `char_col` CHAR(10) COMMENT "", - `decimal_col` decimal COMMENT "" + `decimal_col` decimal COMMENT "", + `json_col` json COMMENT "" ) DISTRIBUTED BY HASH(user_id) PROPERTIES("replication_num" = "1"); """ @@ -85,11 +86,11 @@ suite("test_outfile_csv_with_names") { int i = 1 for (; i < 10; i ++) { sb.append(""" - (${i}, '2017-10-01', '2017-10-01 00:00:00', '2017-10-01', '2017-10-01 00:00:00.111111', '2017-10-01 00:00:00.111111', '2017-10-01 00:00:00.111111', 'Beijing', ${i}, ${i % 128}, true, ${i}, ${i}, ${i}, ${i}.${i}, ${i}.${i}, 'char${i}', ${i}), + (${i}, '2017-10-01', '2017-10-01 00:00:00', '2017-10-01', '2017-10-01 00:00:00.111111', '2017-10-01 00:00:00.111111', '2017-10-01 00:00:00.111111', 'Beijing', ${i}, ${i % 128}, true, ${i}, ${i}, ${i}, ${i}.${i}, ${i}.${i}, 'char${i}', ${i}, '{"a": ${i}, "b": "str${i}"}'), """) } sb.append(""" - (${i}, '2017-10-01', '2017-10-01 00:00:00', '2017-10-01', '2017-10-01 00:00:00.111111', '2017-10-01 00:00:00.111111', '2017-10-01 00:00:00.111111', NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL) + (${i}, '2017-10-01', '2017-10-01 00:00:00', '2017-10-01', '2017-10-01 00:00:00.111111', '2017-10-01 00:00:00.111111', '2017-10-01 00:00:00.111111', NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL) """) sql """ INSERT INTO ${tableName} VALUES ${sb.toString()} @@ -105,13 +106,13 @@ suite("test_outfile_csv_with_names") { } sql """ SELECT * FROM ${tableName} t ORDER BY user_id INTO OUTFILE "file://${outFilePath}/" FORMAT AS CSV_WITH_NAMES - PROPERTIES("column_separator" = ","); + PROPERTIES("column_separator" = "|"); """ File[] files = path.listFiles() assert files.length == 1 // check column names - String columnNames = """user_id,date,datetime,date_1,datetime_1,datetime_2,datetime_3,city,age,sex,bool_col,int_col,bigint_col,largeint_col,float_col,double_col,char_col,decimal_col""" + String columnNames = """user_id|date|datetime|date_1|datetime_1|datetime_2|datetime_3|city|age|sex|bool_col|int_col|bigint_col|largeint_col|float_col|double_col|char_col|decimal_col|json_col""" List<String> outLines = Files.readAllLines(Paths.get(files[0].getAbsolutePath()), StandardCharsets.UTF_8); assertEquals(columnNames, outLines.get(0)) @@ -137,14 +138,15 @@ suite("test_outfile_csv_with_names") { `float_col` float COMMENT "", `double_col` double COMMENT "", `char_col` CHAR(10) COMMENT "", - `decimal_col` decimal COMMENT "" + `decimal_col` decimal COMMENT "", + `json_col` json COMMENT "" ) DISTRIBUTED BY HASH(user_id) PROPERTIES("replication_num" = "1"); """ StringBuilder commandBuilder = new StringBuilder() commandBuilder.append("""curl -v --location-trusted -u ${context.config.feHttpUser}:${context.config.feHttpPassword}""") - commandBuilder.append(""" -H format:csv_with_names -H column_separator:, -T """ + files[0].getAbsolutePath() + """ http://${context.config.feHttpAddress}/api/"""; + dbName + "/" + tableName2 + "/_stream_load") + commandBuilder.append(""" -H format:csv_with_names -H column_separator:| -T """ + files[0].getAbsolutePath() + """ http://${context.config.feHttpAddress}/api/"""; + dbName + "/" + tableName2 + "/_stream_load") command = commandBuilder.toString() process = command.execute() code = process.waitFor() @@ -165,3 +167,4 @@ suite("test_outfile_csv_with_names") { } } } + --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
