嵌套dict和for循环中的python字符串格式

Question

With the following dictionnary, i am trying to implement creating of strings.

query_args= {"columns":{ "investingcom":"datetime, actual, forecast, previous",
                        "bloomberg":"datetime, Actual, `Surv(M)`, Prior",
                        "ft":"datetime, actual, forecast, prior"},
             "table":{"investingcom":"`all-data`.`indicators`",
                     "bloomberg":"`all-data`.`indicators-bloomberg`",
                      "ft":"`all-data`.indicators_ft"},
             "country":{"investingcom":"\"US\"",
                     "bloomberg":"\"US\"",
                      "ft":"\"United States\""},
             "name":{"investingcom":"\"Average Hourly Earnings (MoM)\"",
                     "bloomberg":"\"Average Hourly Earnings MoM\"",
                      "ft":"\"Average Hourly Earnings %% m/m\""}
                     }

There is 3 strings i need to create, hence the use of a for loop:

for ind_source in ["investingcom", "bloomberg", "ft"]:
    print('SELECT {columns[ind_source]} FROM {table[ind_source]} \
    WHERE country={country[ind_source]} AND name={name[ind_source]}'.format(**query_args))

However this returns an error:

KeyError: 'ind_source'

Any suggestion how to make that implementation successful?

Answer 1

I am not familiar with doing it the way you try in your example (though it may be possible to do it that way with some tweaking. However, the way I would do it is something like the following:

query_args= {"columns":{ "investingcom":"datetime, actual, forecast, previous",
                        "bloomberg":"datetime, Actual, `Surv(M)`, Prior",
                        "ft":"datetime, actual, forecast, prior"},
             "table":{"investingcom":"`all-data`.`indicators`",
                     "bloomberg":"`all-data`.`indicators-bloomberg`",
                      "ft":"`all-data`.indicators_ft"},
             "country":{"investingcom":"\"US\"",
                     "bloomberg":"\"US\"",
                      "ft":"\"United States\""},
             "name":{"investingcom":"\"Average Hourly Earnings (MoM)\"",
                     "bloomberg":"\"Average Hourly Earnings MoM\"",
                      "ft":"\"Average Hourly Earnings %% m/m\""}
                     }

for ind_source in ["investingcom", "bloomberg", "ft"]:
    print('SELECT {} FROM {} \
    WHERE country={} AND name={}'.format(
        query_args['columns'][ind_source],
        query_args['table'][ind_source],
        query_args['country'][ind_source],
        query_args['name'][ind_source],
    ))

Answer 2

I would try the following:

for ind_source in ["investingcom", "bloomberg", "ft"]:
     print(
         ("SELECT {columns[%(ind_source)s]} FROM {table[%(ind_source)s]}"
          "WHERE country={country[%(ind_source)s]} AND "
          "name={name[%(ind_source)s]}" % { 'ind_source': ind_source })
     .format(**query_args))

essentially using two format operations in succession to get the queries you want.

Answer 3

How about:

for ind_source in ["investingcom", "bloomberg", "ft"]:
    print(('SELECT {columns[' +ind_source + ']} FROM {table[' + ind_source + ']} \
    WHERE country={country[' + ind_source + ']} AND name={name[' + ind_source + ']}').format(**query_args))

If you put 'ind_source' inside the string, when formatting it will search for the key 'ind_source' in the query_args dictionary. This dictionary doesn't have that key, thus that error. Instead concatenate the string with the current value of ind_source in the for loop.

Answer 4

Instead of trying to do a complicated format, try filtering your data first for a cleaner format string and readability.

query_args= {"columns":{ "investingcom":"datetime, actual, forecast, previous",
                        "bloomberg":"datetime, Actual, `Surv(M)`, Prior",
                        "ft":"datetime, actual, forecast, prior"},
             "table":{"investingcom":"`all-data`.`indicators`",
                     "bloomberg":"`all-data`.`indicators-bloomberg`",
                      "ft":"`all-data`.indicators_ft"},
             "country":{"investingcom":"\"US\"",
                     "bloomberg":"\"US\"",
                      "ft":"\"United States\""},
             "name":{"investingcom":"\"Average Hourly Earnings (MoM)\"",
                     "bloomberg":"\"Average Hourly Earnings MoM\"",
                      "ft":"\"Average Hourly Earnings %% m/m\""}
                     }

for ind_source in ["investingcom", "bloomberg", "ft"]:
    args = {x: query_args[x][ind_source] for x in query_args}
    print('SELECT {columns} FROM {table} \
    WHERE country={country} AND name={name}'.format(**args))