I couldn't find any information about reading ascii tables (prettytable-looking). I need to parse some tables that look like this:

 -------- -------- ------------------- 
|               Planets               |
 -------- -------- ------------------- 
| Planet | R (km) | mass (x 10^29 kg) |
 -------- -------- ------------------- 
|  Sun   | 696000 |    1989100000     |
|(Solar) |        |                   |
 -------- -------- ------------------- 
| Earth  |  6371  |      5973.6       |
 -------- -------- ------------------- 
|  Moon  |  1737  |       73.5        |
 -------- -------- ------------------- 
|  Mars  |  3390  |      641.85       |
 -------- -------- ------------------- 

As you can see, this table contains heading and subtitle, but it isn't a main problem. Here how I tried to parse it (all tries was in Python, maybe exist more suitable language):

• module petl - doesn't work, coud not read -------- -------- -------------------
• np.fromtxt or smth doesnt work too, because there's no array
• module asciitable doesn't work,

asciitable.read("sample.txt",delimiter='|',guess=False,numpy=False, quotechar="'")
InconsistentTableError: Number of header columns (1) inconsistent with data columns (9) at data line 0
Header values: [' ---------- ------------------------------------------------------------ --------------- ----------------- ---------------- -------------------- ----------------------- ']

and of course I tried all the combinations of parametrs

• And yes, tried simple (but not) way and used regular expressions. But in case if "row" have more than 1 line there's too hard to catch all exceptions.
• Also tried simple split, but it unhappy way too...

I heard about numpy substitute by 0 and 1, but its too hard with my table. Please help.

CodePudding user response：

I think you could do this using python, with a few passes you can convert this into something that suits your needs.

table_str = """
 -------- -------- ------------------- 
|               Planets               |
 -------- -------- ------------------- 
| Planet | R (km) | mass (x 10^29 kg) |
 -------- -------- ------------------- 
|  Sun   | 696000 |    1989100000     |
|(Solar) |        |                   |
 -------- -------- ------------------- 
| Earth  |  6371  |      5973.6       |
 -------- -------- ------------------- 
|  Moon  |  1737  |       73.5        |
 -------- -------- ------------------- 
|  Mars  |  3390  |      641.85       |
 -------- -------- ------------------- 
"""

table_list = [
    [item.strip() for item in line.split('|') if item]  # maintain the number of columns in rows.
    for line in table_str.strip().split('\n')
    if ' -' not in line  # discard  -
]

column_count = len(table_list[1])
row_idx_to_remove = list()
for row_idx, row in enumerate(table_list[2:], start=2):
    if any(not c for c in row):  # A multiline column entry
        row_idx_to_remove.append(row_idx)
        for col_idx, col in enumerate(table_list[row_idx - 1]):
            table_list[row_idx - 1][col_idx]  = row[col_idx]

# Remove the rows that have already been merged into previous ones.
for idx in row_idx_to_remove:
    del table_list[idx]

table_dict = {'Name': table_list[0][0], 'Column Names': table_list[1], 'Rows': table_list[2:]}

print(table_dict)
print(table_list)

Output:

{'Name': 'Planets', 'Column Names': ['Planet', 'R (km)', 'mass (x 10^29 kg)'], 'Rows': [['Sun(Solar)', '696000', '1989100000'], ['Earth', '6371', '5973.6'], ['Moon', '1737', '73.5'], ['Mars', '3390', '641.85']]}
[['Planets'], ['Planet', 'R (km)', 'mass (x 10^29 kg)'], ['Sun(Solar)', '696000', '1989100000'], ['Earth', '6371', '5973.6'], ['Moon', '1737', '73.5'], ['Mars', '3390', '641.85']]

This implementation does not handle empty columns in rows nor does it handle the case where every column in a row has more than one line.

You could also make it use re.split, and split on lines containing -, then you could handle any rows that would contain the edge cases of this implementation.