FAQ

Table Conversion Issues

When will the output_censored directory be generated?

• When generating a single table, if --censored-fields <fields> is passed, both censored version and complete version output files will be generated, corresponding to the output_censored/ and output/ directories respectively.

• When batch generating tables, if censoredTable or censoredFields fields are configured in $tables.xlsx, both censored version and complete version output files will be generated. Tables marked with censoredTable will not be output to the output_censored/ directory, only to output/.

• If censoredTable or censoredFields fields are not configured, only one output/ will be generated.

About file_identifier

When using x2f, the .fbs file content’s SHA-256 hash is calculated, and the first four characters are used as the file_identifier.

You can use the BufferHasIdentifier method in the generated code to verify whether the binary file matches the code. (If you are using a Unity Loader class generated by x2f, the Load function will automatically perform this validation.)

uint64/int64 Precision Issues

When storing numbers like 9007199254740993 in the table, precision may be lost. You can set the cell to text format to preserve precision.

Runtime Issues

[ C# ] System.OverflowException: Value was either too large or too small for an Int16.

• Trigger condition: When verifying the buffer, for example:

  Xls.SomeTable.Instance.LoadAsync();
  Xls.MergeTableLoader.LoadAllAsync();

• Cause: When there is a large amount of data in the table, the offset exceeds the range of the short type in the verification method.

  FlatBuffers/FlatBufferVerify.cs

  private short GetVRelOffset(int pos, short vtableOffset)
  {
      short VOffset = 0;
      // Used try/catch because pos typa as int 32bit
      try
      {
      // First, get vtable offset
      short vtable = Convert.ToInt16(pos - ReadSOffsetT(verifier_buffer, pos));
      // Check that offset points to vtable area (is smaller than vtable size)
      if (vtableOffset < ReadVOffsetT(verifier_buffer, vtable))
      {
          // Now, we can read offset value - TODO check this value against size of table data
          VOffset = ReadVOffsetT(verifier_buffer, vtable + vtableOffset);
      }
      // (...)
  }

• Solutions:

  1. Modify the GetVRelOffset method in FlatBuffers to use int type instead of short type.
  FlatBuffers/FlatBufferVerify.cs

  // short vtable = Convert.ToInt16(pos - ReadSOffsetT(verifier_buffer, pos));
  // 👇 Change to this
  int vtable = pos - ReadSOffsetT(verifier_buffer, pos);
  2. Or modify the TableValidator.Validate method in UnityTemplate to only verify that the identifier matches, without verifying the buffer structure.