def get_transform():
w = dw.DataWrangler()
# Split data repeatedly on newline into rows
w.add(
dw.Split(column=["data"],
table=0,
status="active",
drop=True,
result="row",
update=False,
insert_position="right",
row=None,
on="\n",
before=None,
after=None,
ignore_between=None,
which=1,
max=0,
positions=None,
quote_character=None))
# Delete empty rows
w.add(
dw.Filter(column=[],
table=0,
status="active",
drop=False,
row=dw.Row(column=[],
table=0,
status="active",
drop=False,
conditions=[
dw.Empty(column=[],
table=0,
status="active",
drop=False,
percent_valid=0,
num_valid=0)
])))
# Delete rows where data starts with '==='
w.add(
dw.Filter(column=[],
table=0,
status="active",
drop=False,
row=dw.Row(column=[],
table=0,
status="active",
drop=False,
conditions=[
dw.StartsWith(column=[],
table=0,
status="active",
drop=False,
lcol="data",
value="===",
op_str="starts with")
])))
# Delete rows where data = '<!-- KBDX was Broadus Airport ...
w.add(
dw.Filter(
column=[],
table=0,
status="active",
drop=False,
row=dw.Row(
column=[],
table=0,
status="active",
drop=False,
conditions=[
dw.
Eq(column=[],
table=0,
status="active",
drop=False,
lcol="data",
value=
"<!-- KBDX was Broadus Airport in Broadus, Montana. Replaced by new airport with FAA ID: 00F -->",
op_str="=")
])))
# Delete rows where data contains '<s>''''
w.add(
dw.Filter(column=[],
table=0,
status="active",
drop=False,
row=dw.Row(column=[],
table=0,
status="active",
drop=False,
conditions=[
dw.Contains(column=[],
table=0,
status="active",
drop=False,
lcol="data",
value="<s>'''",
op_str="contains")
])))
# Extract from data between positions 5, 9
w.add(
dw.Extract(column=["data"],
table=0,
status="active",
drop=False,
result="column",
update=False,
insert_position="right",
row=None,
on=None,
before=None,
after=None,
ignore_between=None,
which=1,
max=1,
positions=[5, 9]))
# Drop data
w.add(dw.Drop(column=["data"], table=0, status="active", drop=True))
return w
from wrangler import dw
w = dw.DataWrangler()
# Split data repeatedly on newline into rows
w.add(dw.Split(column=["data"],
table=0,
status="active",
drop=True,
result="row",
update=False,
insert_position="right",
row=None,
on="\n",
before=None,
after=None,
ignore_between=None,
which=1,
max=0,
positions=None,
quote_character=None))
# Split data repeatedly on ','
w.add(dw.Split(column=["data"],
table=0,
status="active",
drop=True,
result="column",
update=False,
insert_position="right",
row=None,
sys.exit(
'Error: Please include an input and output file. Example python script.py input.csv output.csv'
)
w = dw.DataWrangler()
# Split data repeatedly on '|-' into rows
w.add(
dw.Split(column=["data"],
table=0,
status="active",
drop=True,
result="row",
update=False,
insert_position="right",
row=None,
on="\\|-",
before=None,
after=None,
ignore_between=None,
which=1,
max=0,
positions=None,
quote_character=None))
# Cut from data on '| any lowercase word =#FFF any number any word \|'
w.add(
dw.Cut(column=["data"],
table=0,
status="active",
drop=False,
result="column",
