Keep the original class of a value imported from SQL to R

Question

I imported a table from an SQL database, but R is automatically converting the values from a column of character for num and I need him to preserve them in class character. It seems that this occurs because the package RODBC behaves similar to Excel.

Table data to be imported:

TABLE_CAT TABLE_SCHEM           TABLE_NAME             COLUMN_NAME DATA_TYPE TYPE_NAME COLUMN_SIZE BUFFER_LENGTH
5  BI_TOXICOLOGIA         dbo TA_HISTORICO_PETICAO             NU_PROCESSO        12   varchar          21            21
6  BI_TOXICOLOGIA         dbo TA_HISTORICO_PETICAO           NU_EXPEDIENTE        12   varchar          21            21

The columns NU_PROCESSO and NU_EXPEDIENTE, are numeric strings, but need to continue in format CHAR.

I used the package RODBC to import the table:

con <- odbcConnect("BISQL", uid = "login", pwd = "senha")
ANVS_base_completa_temp <- sqlFetch(con, "TA_HISTORICO_PETICAO")

In the result the columns are coming as NUM and end up losing data integrity. Even if I convert them again to character.

str(ANVS_base_completa_temp)
'data.frame':   193137 obs. of  15 variables:
 $ NU_PROCESSO            : num  2.54e+16 2.54e+16 2.54e+16 2.54e+16 2.54e+16 ...
 $ NU_EXPEDIENTE          : num  1.47e+08 1.38e+08 1.26e+08 1.13e+08 7.79e+07 ...

Does anyone know how I can import this table without R converting this data? Suddenly using some other package it works?