Iterating over table cells in order.

[_savage]

I’d like to visit all cells of a table in order, left-to-right (row) and top-to-bottom would be fine. I’ve read this wiki page but would like some clarification.

Suppose I iterate over the rows (interface XTableRows, service TextTableRow):

Code: Select all

for row in table.getRows():                                             
    print("row " + str(row) + " cells " + str(row.getCount()))

Actually gives me an error: AttributeError: getCount. Inspecting the same document with MRI shows that there is indeed a Count attribute and a getCount() method for row, yet the AttributeError exception. Either way, having a row still would not give me access to the cells of that row, would it?

Next I tried to iterate over all cells by their name (method getCellNames() of the XTextTable interface):

Code: Select all

for n in table.getCellNames():                                        
    cell = table.getCellByName(n)

which works and seems to iterate in-order over all rows top down. However, can I assume that the sequence of cell names is always in that same order?

Furthermore, looking and the cell names parameter to getCellByName() it seems that table cell names follow a naming convention: columns are enumerated with capital letters A ... Z, and rows with numbers. How does that scheme extend to larger tables, will columns then be named AA, AB, AC, ... , AZ, BA, BB, BC, ... and so forth?

[F3K Total]

If the tables name is Table1, this should help.

Code: Select all

Sub iterate_TextTableCells
    oTexttables = ThisComponent.Texttables
    oTexttable = oTexttables.getByName("Table1")
    nRows = oTexttable.Rows.count - 1
    nColumns = oTexttable.Columns.count - 1
    for i = 0 to nRows
        for k = 0 to nColumns
            oCell = oTexttable.getCellbyPosition(k,i)
            Print "Cell: " & ocell.CellName & " Value: " & ocell.Value 'for numbers, .string for text
        next k
    next i
End Sub

[_savage]

Thank you, that seems to work for simple tables. However, how does this approach iterate over split or spanning cells? For example, for this table:

Code: Select all

+----+----+----+-----+
|    |    |    |     |  (4 columns)
+----+----+----+-----+
|    |               |  (2 columns)
+----+---+---+---+---+
|    |   |   |   |   |  (5 columns)
+----+---+---+---+---+

the above code snippet gives me 3 rows (ok) and 4 columns (wrong), which causes an OutOfBounds in row 2 when I read column 3 or 4, and misses column 5 in row 3.

[F3K Total]

You're right. Doesn't work for split or spanning cells, use your 2nd approach.
Texttables seem to use the alphabetical order, but also lower case for the columns:
A,B...Z
a,b...z
AA,AB ...AZ
Aa,Ab ...Az...
R

[_savage]

You're right. Doesn't work for split or spanning cells, use your 2nd approach.

So my worry is that getCellNames() returns an unordered sequence of cell names, which would lead to a more or less random visitation of the table’s cell. I can’t find documented that this functions returns an ordered sequence of cell names...

[F3K Total]

Try it

[F3K Total]

Here i colored them, you can see that everything works fine first from left to right then from top to bottom.

[_savage]

The thing that irritates me, and looking at this from a HTML colspan, rowspan perspective, is that I find it difficult to find out whether a cell spans others or if two others are a split cell.

For example, in the above table all cells are equal simple cells, yet one can argue that they split and span: B2 spans A2, A3, A4; or C2, C3, C4, C5 split B2 but none of the A2/3/4 spans any of the C2/3/4/5 cells.

Or another example:

Code: Select all

+---------+
|         |  <- Does this span the bottom two cells, or
+----+----+
|    |    |  <- is this a split cell?
+----+----+

Finding out if a cell is indeed a split cell can be done by its name. But then I still don’t really know which cells in relation it splits... Perhaps Office table setup isn't as complex and flexible as HTML?

[F3K Total]

Yes, you're right, i can't help further, do also find no information but only the TableColumnSeparators of the rows, maybe the position helps you.