Method ExtractFields

tensorflow/core/kernels/decode_csv_op.cc:218–303

  string na_value_;

  void ExtractFields(OpKernelContext* ctx, StringPiece input,
                     std::vector<string>* result) {
    int64 current_idx = 0;
    int64 num_fields_parsed = 0;
    int64 selector_idx = 0;  // Keep track of index into select_cols

    if (!input.empty()) {
      while (static_cast<size_t>(current_idx) < input.size()) {
        if (input[current_idx] == '\n' || input[current_idx] == '\r') {
          current_idx++;
          continue;
        }

        bool quoted = false;
        bool include =
            (select_all_cols_ || select_cols_[selector_idx] ==
                                     static_cast<size_t>(num_fields_parsed));

        if (use_quote_delim_ && input[current_idx] == '"') {
          quoted = true;
          current_idx++;
        }

        // This is the body of the field;
        string field;
        if (!quoted) {
          while (static_cast<size_t>(current_idx) < input.size() &&
                 input[current_idx] != delim_) {
            OP_REQUIRES(ctx,
                        (!use_quote_delim_ || input[current_idx] != '"') &&
                            input[current_idx] != '\n' &&
                            input[current_idx] != '\r',
                        errors::InvalidArgument(
                            "Unquoted fields cannot have quotes/CRLFs inside"));
            if (include) field += input[current_idx];
            current_idx++;
          }

          // Go to next field or the end
          current_idx++;
        } else if (use_quote_delim_) {
          // Quoted field needs to be ended with '"' and delim or end
          while (
              (static_cast<size_t>(current_idx) < input.size() - 1) &&
              (input[current_idx] != '"' || input[current_idx + 1] != delim_)) {
            if (input[current_idx] != '"') {
              if (include) field += input[current_idx];
              current_idx++;
            } else {
              OP_REQUIRES(
                  ctx, input[current_idx + 1] == '"',
                  errors::InvalidArgument("Quote inside a string has to be "
                                          "escaped by another quote"));
              if (include) field += '"';
              current_idx += 2;
            }
          }