我有一个数据框：
df
Out[20]:
StreetAddressLine1 StateAbbreviation ... Longitude BetId
0 Unknown Unknown ... None 0
1 21236 Birchwood Loop AK ... None 1
2 1731 Bragaw St AK ... None 2
3 4360 Snider Dr AK ... None 4
4 9750 W Parks Hwy AK ... None 10
5 7205 Shorewood Dr AK ... None 11
6 326 Woodside Ave AK ... None 14
7 2036 E Northern Lights Blvd AK ... None 15
8 1600 E Tudor Rd AK ... None 16
9 1130, 2545 E Tudor Rd AK ... None 17
我运行我的代码对这些地址进行地理编码：
input_file_path = "df"
output_file_path = "output" # appends "####.csv" to the file name when it writes the file.
# Set the name of the column indexes here so that pandas can read the CSV file
address_column_name = "StreetAddressLine1"
state_column_name = "StateAbbreviation"
zip_column_name = "ZipCode" # Leave blank("") if you do not have zip codes
# Where the program starts processing the addresses in the input file
# This is useful in case the computer crashes so you can resume the program where it left off or so you can run multiple
# instances of the program starting at different spots in the input file
start_index = 0
# How often the program prints the status of the running program
status_rate = 100
# How often the program saves a backup file
write_data_rate = 1000
# How many times the program tries to geocode an address before it gives up
attempts_to_geocode = 3
# Time it delays each time it does not find an address
# Note that this is added to itself each time it fails so it should not be set to a large number
wait_time = 3
# ----------------------------- Processing the input file -----------------------------#
#df = pd.read_csv(input_file_path, low_memory=False,encoding="utf-8")
# df = pd.read_excel(input_file_path)
# Raise errors if the provided column names could not be found in the input file
if address_column_name not in df.columns:
raise ValueError("Can't find the address column in the input file.")
if state_column_name not in df.columns:
raise ValueError("Can't find the state column in the input file.")
# Zip code is not needed but helps provide more accurate locations
if (zip_column_name):
if zip_column_name not in df.columns:
raise ValueError("Can't find the zip code column in the input file.")
addresses = (df[address_column_name] + ', ' + df[zip_column_name].astype(str) + ', ' + df[state_column_name]).tolist()
else:
addresses = (df[address_column_name] + ', ' + df[state_column_name]).tolist()
# ----------------------------- Function Definitions -----------------------------#
# Creates request sessions for geocoding
class GeoSessions:
def __init__(self):
self.Arcgis = requests.Session()
self.Komoot = requests.Session()
# Class that is used to return 3 new sessions for each geocoding source
def create_sessions():
return GeoSessions()
# Main geocoding function that uses the geocoding package to covert addresses into lat, longs
def geocode_address(address, s):
g = geocoder.arcgis(address, session=s.Arcgis)
if (g.ok == False):
g = geocoder.komoot(address, session=s.Komoot)
return g
def try_address(address, s, attempts_remaining, wait_time):
g = geocode_address(address, s)
if (g.ok == False):
time.sleep(wait_time)
s = create_sessions() # It is not very likely that we can't find an address so we create new sessions and wait
if (attempts_remaining > 0):
try_address(address, s, attempts_remaining-1, wait_time+wait_time)
return g
# Function used to write data to the output file
def write_data(data, index):
file_name = (output_file_path + str(index) + ".csv")
print("Created the file: " + file_name)
done = pd.DataFrame(data)
done.columns = ['Address', 'Lat', 'Long']
done.to_csv((file_name + ".csv"), sep=',', encoding='utf8')
# Variables used in the main for loop that do not need to be modified
s = create_sessions()
results = []
failed = 0
total_failed = 0
progress = len(addresses) - start_index
# ----------------------------- Main Loop -----------------------------#
for i, address in enumerate(addresses[start_index:]):
# Print the status of how many addresses have be processed so far and how many of the failed.
if ((start_index + i) % status_rate == 0):
total_failed += failed
print(
"Completed {} of {}. Failed {} for this section and {} in total.".format(i + start_index, progress, failed,
total_failed))
failed = 0
# Try geocoding the addresses
try:
g = try_address(address, s, attempts_to_geocode, wait_time)
if (g.ok == False):
results.append([address, "was", "not", "geocoded"])
print("Gave up on address: " + address)
failed += 1
else:
results.append([address, g.latlng[0], g.latlng[1]])
# If we failed with an error like a timeout we will try the address again after we wait 5 secs
except Exception as e:
print("Failed with error {} on address {}. Will try again.".format(e, address))
try:
time.sleep(5)
s = create_sessions()
g = geocode_address(address, s)
if (g.ok == False):
print("Did not fine it.")
results.append([address, "was", "not", "geocoded"])
failed += 1
else:
print("Successfully found it.")
results.append([address, g.latlng[0], g.latlng[1]])
except Exception as e:
print("Failed with error {} on address {} again.".format(e, address))
failed += 1
results.append([address, e, e, "ERROR"])
# Writing what has been processed so far to an output file
if (i%write_data_rate == 0 and i != 0):
write_data(results, i + start_index)
# print(i, g.latlng, g.provider)
# Finished
write_data(results, i + start_index + 1)
print("Finished! :)")
我希望最终输出文件中的结果也能反映这些位置的BetId。
我尝试了
results.append(df["BetId"])
结果
Out[40]:
[['Unknown, Unknown, Unknown', 25.851060000000075, 88.24131000000006],
['21236 Birchwood Loop, 99567, AK', 61.408868754342635, -149.48655639165537],
['1731 Bragaw St, 99508, AK', 61.204894742714515, -149.80829304403093],
['4360 Snider Dr, 99654, AK', 61.58477348398676, -149.34070982806276],
['9750 W Parks Hwy, 99652, AK', 61.56803449899039, -149.69619047155058],
['7205 Shorewood Dr, 99645, AK', 61.626084461047675, -149.2686871507012],
['326 Woodside Ave, 99603, AK', 59.64314849321932, -151.55260136081137],
['2036 E Northern Lights Blvd, 99508, AK',
61.19525951731225,
-149.8425921931733],
['1600 E Tudor Rd, 99507, AK', 61.180762485140605, -149.851269],
['1130, 2545 E Tudor Rd, 99507, AK', 61.180918519331485, -149.83301780682672],
0 0
1 1
2 2
3 4
4 10
5 11
6 14
7 15
8 16
9 17
Name: BetId, dtype: int64]
但是您可以看到BetId不在lat和long之后的第4列附加。
我也尝试过
write_data(results.append(df["BetId"]), i + start_index + 1)
但是我收到一个错误：
ValueError：长度不匹配：预期轴包含0个元素，新值包含3个元素
如何解决此问题，以便最终输出的csv反映除地理编码以外的原始数据帧中的下注ID
答案 0 :(得分：1)
现在，您要将Pandas系列附加到一个标准的Python列表中，即两个不同的对象。由于 BetId 与 addresses 的长度相同，因为二者均源自与列相同的数据帧，因此请使用枚举循环变量 i 来索引< em> BetId ，然后将值添加为列表的第4个元素。在处理结果期间执行此操作，不要在以下时间进行
：for i, address in enumerate(addresses[start_index:]):
...
results.append([address, g.latlng[0], g.latlng[1], df["BetId"].loc[i]])