Send quoting settings with CsvSettings (ydb-platform#4210)

maximyurchuk · May 1, 2024 · c4610d9 · c4610d9
1 parent e0cc2ab
commit c4610d9
Show file tree

Hide file tree

Showing 2 changed files with 14 additions and 0 deletions.
diff --git a/ydb/core/grpc_services/rpc_load_rows.cpp b/ydb/core/grpc_services/rpc_load_rows.cpp
@@ -467,6 +467,13 @@ class TUploadColumnsRPCPublic : public NTxProxy::TUploadRowsBase<NKikimrServices
                     errorMessage = reader.status().ToString();
                     return false;
                 }
+                const auto& quoting = cvsSettings.quoting();
+                if (quoting.quote_char().length() > 1) {
+                    errorMessage = TStringBuilder() << "Wrong quote char '" << quoting.quote_char() << "'";
+                    return false;
+                }
+                const char qchar = quoting.quote_char().empty() ? '"' : quoting.quote_char().front();
+                reader->SetQuoting(!quoting.disabled(), qchar, !quoting.double_quote_disabled());
                 reader->SetSkipRows(skipRows);
 
                 if (!delimiter.empty()) {

diff --git a/ydb/public/api/protos/ydb_formats.proto b/ydb/public/api/protos/ydb_formats.proto
@@ -9,6 +9,12 @@ message ArrowBatchSettings {
 }
 
 message CsvSettings {
+    message Quoting {
+        bool disabled = 1;
+        bytes quote_char = 2;
+        bool double_quote_disabled = 3;
+    }
+
     // Number of rows to skip before CSV data. It should be present only in the first upsert of CSV file.
     uint32 skip_rows = 1;
     // Fields delimiter in CSV file. It's "," if not set.
@@ -17,4 +23,5 @@ message CsvSettings {
     bytes null_value = 3;
     // First not skipped line is a CSV header (list of column names).
     bool header = 4;
+    Quoting quoting = 5;
 }