Question

使用C＃中的HDF.PInvoke库，如何从可变长度字符串H5T_VARIABLE属性中读取内容？

HDF5文件如下：

GROUP "/" {
   ATTRIBUTE "foo" {
      DATATYPE  H5T_STRING {
         STRSIZE H5T_VARIABLE;
         STRPAD H5T_STR_NULLTERM;
         CSET H5T_CSET_ASCII;
         CTYPE H5T_C_S1;
      }
      DATASPACE  SCALAR
      DATA {
      (0): "bar"
      }
   }
}

尝试像这样读取属性foo：

long fileId = H5F.open("my-file.h5", H5F.ACC_RDWR);
long attrId = H5A.open(fileId, "foo");
long typeId = H5A.get_type(attrId);
H5A.info_t attrInfo = new H5A.info_t();
var info_result = H5A.get_info(attrId, ref attrInfo);

// Note sure if this `size` is useful.
// Docs say: For variable-length string datatypes, the returned value is
//           the size of the pointer to the actual string, or sizeof(char *). 
//           This function does not return the size of actual variable-length
//           string data.
// @See: https://support.hdfgroup.org/HDF5/doc/RM/RM_H5T.html#Datatype-GetSize
int size = H5T.get_size(typeId).ToInt32();

// Docs say: `data_size` indicates the size, in the number of characters,
//           of the attribute.
// @See: https://support.hdfgroup.org/HDF5/doc/RM/RM_H5A.html#Annot-GetInfo
int data_size = (int)attrInfo.data_size;

// Docs say: In a C environment, variable-length strings will always be
//           NULL-terminated, so the buffer to hold such a string must be
//           one byte larger than the string itself to accommodate the NULL
//           terminator.
// @See: https://support.hdfgroup.org/HDF5/doc/RM/RM_H5T.html#CreateVLString
IntPtr iPtr = Marshal.AllocHGlobal(data_size + 1); // Add one to fit NULL-terminator.

int read_result = H5A.read(attrId, typeId, iPtr);
if (H5T.is_variable_str(typeId) > 0)
{
  Console.WriteLine("attribute value = {0}", Marshal.PtrToStringAnsi(iPtr));
}

控制台输出随机数据：

attribute value = @Ñ>←

Answer 1

看看HDFql，因为在C＃中处理HDF可变长度字符串时，这将使您摆脱底层细节的困扰。下面是一个完整的示例，说明如何使用HDFql读取（和打印）存储在属性中的可变长度字符串（假设存在名为_fopen的HDF文件，并且其中包含可变长度的属性my-file.h5字符串）：

foo

Answer 2

我从Hdf5DotnetTools库中找到了一个解决方案。

long fileId = H5F.open("my-file.h5", H5F.ACC_RDWR);
long attrId = H5A.open(fileId, "foo");
long typeId = H5A.get_type(attrId);
long spaceId = H5A.get_space(attrId);
long count = H5S.get_simple_extent_npoints(spaceId);
H5S.close(spaceId);

IntPtr[] dest = new IntPtr[count];
GCHandle handle = GCHandle.Alloc(dest, GCHandleType.Pinned);
H5A.read(attrId, typeId, handle.AddrOfPinnedObject());

var attrStrings = new List<string>();
for (int i = 0; i < dest.Length; ++i)
{
    int attrLength = 0;
    while (Marshal.ReadByte(dest[i], attrLength) != 0)
    {
        ++attrLength;
    }

    byte[] buffer = new byte[attrLength];
    Marshal.Copy(dest[i], buffer, 0, buffer.Length);
    string stringPart = Encoding.UTF8.GetString(buffer);

    attrStrings.Add(stringPart);

    H5.free_memory(dest[i]);
}

handle.Free();

if (H5T.is_variable_str(typeId) > 0)
{
    Console.WriteLine("attribute value = {0}", attrStrings[0]);
}

控制台输出正确：

attribute value = bar

如何在C＃.NET中读取HDF5可变长度字符串属性

2 个答案: