How does Pandas allocate memory for pooling / pooling internally?

How do Pandas internal functions allocate memory for operations on an indefinite number of output lines? For example, when concatenating two data frames, the number of output lines depends on the values ​​of the inputs. You don't really know how many lines there will be until after you finish the connection. Does Pandas do two passes? Does it use a dynamic data structure? If so, what?