Student Recreation of "Colossus"
Student Recreation of "Colossus"
Student Recreation of "Colossus"

Panda3D Manual: More about GeomVertexReader, GeomVertexWriter, and GeomVertexRewriter

The classes GeomVertexReader and GeomVertexWriter together represent the core interface for reading and writing the numeric data stored within a GeomVertexData object.

These two classes work similarly. Both are designed to get a temporary pointer to the data for a particular column when they are constructed, and they increment that pointer as you walk through the vertices. Initially, they start at row number 0 (the first vertex in the table), and after each setData/getData operation, they automatically increment to the next row (the next vertex).

You construct a GeomVertexReader or GeomVertexWriter with a pointer to the GeomVertexData object you are operating on, and the name of the column you wish to process, e.g.:

color = GeomVertexReader(vdata, 'color')

Because the GeomVertexReader and GeomVertexWriter classes only store a temporary pointer, which might become invalid between frames or even between different tasks within a frame, these objects should not be stored in any persistent object. Instead, they are designed to be temporary objects that are constructed locally, used immediately to iterate through a list of vertices, and then released. If you need to keep a persistent iterator for your vertex data, to be used over a long period of time (e.g. over several frames), then you should store just the GeomVertexData pointer (along with the current vertex index number if you require this), and construct a temporary GeomVertexReader/Writer each time you need to access it.

The following methods are available to read and write data in a column:

GeomVertexReader GeomVertexWriter
x = getData1f() setData1f(x)addData1f(x)
v2 = getData2f() setData2f(x, y)setData2f(v2)addData2f(x, y)addData2f(v2)
v3 = getData3f() setData3f(x, y, z)setData3f(v3)addData3f(x, y, z)addData3f(v3)
v4 = getData4f() setData4f(x, y, z, w)setData4f(v4)addData4f(x, y, z, w)addData4f(v4)
x = getData1i() setData1i(x)addData1i(x)
setData2i(x, y)addData2i(x, y)
setData3i(x, y, z)addData3i(x, y, z)
setData4i(x, y, z, w)addData4i(x, y, z, w)

Each of the getData family of functions supported by GeomVertexReader returns the value of the data in the current column, converted to the requested type. The 'f' suffix indicates a floating-point value, while 'i' indicates an integer value; the digit indicates the number of components you expect to receive.

For instance, getData2f() always returns a VBase2, regardless of the type of data actually stored in the column. If the column contains a 2-component value such as a 2-D texture coordinate, then the returned value will represent the (U, V) value in that column. However, if the column type does not match the requested type, a conversion is quietly made; for instance, if you call getData2f() but the column actually contains a 3-D texture coordinate, the third component will be omitted from the return value, which will still be a VBase2.

Similarly, the setData and addData family of functions supported by GeomVertexWriter accept a value in the indicated format, and convert it to whatever format is required by the column. So if you call setData3f(), and the column has three components, you will set all three components with the x, y, z parameters of setData3f(); but if the column has only two components, only the x, y parameters will be used to set those two components, and the third parameter will be ignored.

Certain kinds of numeric conversions are performed automatically, according to the column's designated contents. For instance, if you store a floating-point value into an integer column, the fractional part of the value is usually truncated. However, if the column contents indicates that it represents a color value, then the floating-point value is automatically scaled from the range 0.0 .. 1.0 into the full numeric range of the column's integer value. This allows you to store color components in the range 0.0 .. 1.0, and get the expected result (that is, the value is scaled into the range 0 .. 255). A similar conversion happens when data is read.

There are no getData2i, 3i, or 4i methods available, simply because Panda does not currently define a multi-component integer value that can be returned to Python. Since most multi-component column types are floating-point, or can be expressed as floating-point, this is not generally a limitation.

Each GeomVertexReader keeps track of the current read row, which is initially 0; the current value can be retrieved by getReadRow(). Each call to a getData function returns the value of the column at the current read row, and then increments the current read row. It is an error to call getData when the read row has reached the end of the data, but you can call isAtEnd(), which returns true when the reader has reached the end. Thus, you can iterate through all the rows of a vertex table by repeatedly calling getData until isAtEnd() returns true.

Similarly, each GeomVertexWriter keeps track of the current write row, which is initially 0, and can be retrieved by getWriteRow(). Each call to setData or addData stores the given value in the current write row, and then increments the current write row. It is an error to call setData when the write row has reached the end of the data; but as with the GeomVertexReader, you can call isAtEnd() to determine when you have reached the end of the data.

The addData family of functions work exactly like the setData functions, except that addData can be called when the GeomVertexWriter has reached its end. In this case, addData will add a new row to the table, and then fill in the specified data in that row (and then increment the current write row). If addData is called when the current write row already exists, it behaves exactly the same as setData.

With either GeomVertexReader or GeomVertexWriter, you can set the current read or write row at any time with the call setRow(). This sets the current read row (GeomVertexReader) or current write row (GeomVertexWriter) to the indicated value; the next call to getData or setData/addData will then operate on the specified row, and increment from there.


The GeomVertexRewriter class exists as a convenience for code that needs to alternately read and write the data on a column. GeomVertexRewriter multiply inherits from GeomVertexReader and GeomVertexWriter, so it supports the getData family of functions, as well as the setData and addData family of functions. It also has both a current read row and a current write row, which might be different.

Normally, you would use a GeomVertexRewriter to walk through the list of vertices from the beginning to end, reading and writing as it goes. For instance, to set all of the Z components of a piece of geometry to 0.0, while preserving the X and Y components, you might write a loop such as:

vertex = GeomVertexRewriter(vdata, 'vertex')
while not vertex.isAtEnd():
v = vertex.getData3f()
vertex.setData3f(v[0], v[1], 0.0)

Note that this example code calls getData3f() and setData3f() exactly once through each iteration, which increments the current read row and current write row, respectively; so the current read row and current write row are kept in sync with each other.

Important! When you are simultaneously reading from and writing to the same GeomVertexData object, you should create all of the GeomVertexWriters and GeomVertexRewriters you need before you create any GeomVertexReader. This is because of Panda's internal referencing-counting mechanism; creating a GeomVertexWriter may automatically (and transparently) force a copy of the data in the GeomVertexData, which could invalidate any GeomVertexReaders you have already created.