multiple imputation methods: ice running error

[Anna Novaresio]

Dear all,

I am trying to imputing missing data using the following command:
ice BROWN_TECHS_POP GREEN_TECHS_POP EPS DIESEL_TAX NOX_TAX NOX_EMISSION_LIMIT PM_EMISSION_LIMIT SULPHUR_CONTENT_DIESEL_LIMIT TOT_PAT ENVI_PAT RATIO_ENVITOT_PAT RD_TOT GREEN_RD FOSSILFUEL_RD CAR_PROD CAR_SALES BEV_SALES FUEL_PRICE GDP_PC GDP POP, saving(icedata) m(10) seed(123)

but I get the following error message;
Perfect prediction detected: using augmlogit to impute NOX_EMISSION_LIMIT
Error #430 encountered while running -uvis-..."

I tried to impute only the variables for which Stata detected a perfect prediction, that are NOX_EMISSION_LIMIT PM_EMISSION_LIMIT SULPHUR_CONTENT_DIESEL_LIMIT, and it tells that they appear as 'complete', while they aren't (I know it since I have seen in the dataset that they have missing values as NOX_TAX, etc, also with similar pattern).
If I discard/delete them from my ice command, my imputed dataset shows missing observations as the original one and it is useless for my purpose.

Could you help me, please, to run a complete imputation on my dataset, with all the imputed/above-mentioned variables?
Many thanks in advance!
Anna

PS. this kind of error occurs to me only with the current dataset, while not running the ice command with other datasets (what's wrong with those variables which are almost identical to other in the dataset, e.g. NO_TAX, etc?)

[Rich Goldstein]

your issue is not completely clear to me, but I suggest adding the "persist" option to your command

[Anna Novaresio]

it works by adding "persist", thank you a lot, Rich Goldstein !

[Anna Novaresio]

PS Using "force", as I did before, it wasn't producing the intended result.