to_xml and read_xml

[hayd]

http://stackoverflow.com/questions/18574108/how-do-convert-a-pandas-dataframe-to-xml
cc @jreback

I don't think there is a standard xml, but perhaps his could be generalised? If not maybe simply adding this recipe to the cookbook is useful.

Solutions for read_xml are hacking tree with from_records
http://stackoverflow.com/a/10948591/1240268
or just as a list http://stackoverflow.com/a/16993660/1240268

too much of a dark art to generalise?

[cpcloud]

could potentially base read_html on this

[cpcloud]

Could have elements like: value (values in a frame), column (column values), index (index values)

where a frame like

In [10]: df = DataFrame({'a':[1,2],'b':[3,4]})

In [11]: df.index.name = 'bob'; df.columns.name = 'jones'

In [12]: df
Out[12]:
jones  a  b
bob
0      1  3
1      2  4

results in the following XML

In [13]: print(df.to_xml())
<frame>
    <names>
      <axis value="0">bob</axis>
      <axis value="1">jones</axis>
    <names>
    <column name="a">
      <value index="0">1</value>
      <value index="1">2</value>
    </column>
    <column name="b">
      <value index="0">3</value>
      <value index="1">4</value>
    </column>
</frame>

For MultiIndex, axis elements could optionally take a level="the_level" attribute.

[jtratner]

I feel like this is more useful for a specific case like 'convert this into
something that XYZ can read and happens to be xml' (which is basically what
the excel funcs do). Do we have a few concrete use case or cases to
discuss?