Wrong table name in scope response

Question

My contract has two tables mcat and mcatmig.

Problem: scope request return wrong table name mcat........1 instead of mcatmig:

% cleos get scope teresacode
{
"rows": [{
    "code": "teresacode",
    "scope": "default",
    "table": "mcat",
    "payer": "teresacode",
    "count": 5688
    },{
    "code": "teresacode",
    "scope": "default",
    "table": "mcat........1",
    "payer": "teresacode",
    "count": 2844
    }
],
"more": ""
}

And there is no table with that name:

% cleos get table teresacode default "mcat........1"
Error 3060003: Contract Table Query Exception
Most likely, the given table doesn't exist in the blockchain.
Error Details:
Unsupported table name: mcat........1

And real table is accessible fine:

% cleos get table teresacode default "mcatmig"
{
"rows": [],
"more": false,
"next_key": ""
}

I believe this happened after copying a large number of rows from table mcat to table mcatmig.

Something like that:

[[eosio::action]] void mcat::move(int count) {
    mcat_t mcat{get_self(), name{"default"}};
    mcatmig_t mcatmig{get_self(), name{"default"}};

    auto it = mcat.begin();
    int i = 0;
    while(i < count && it != mcat.end()) {
        mcatmig.emplace(...);
        i++;
        mcat.erase(it);
        it = mcat.begin();
    }
}

Why scope request return weird table name? Is this a data corruption, move algorithm has a bug? Or it is an eosio bug? What do you think?

Thank you!

Answer 1

If your multi_index table has multiple indices by using indexed_by, additional index will be shown like ORIGINAL_TABLE_NAME........1. The last character will use remaining last 4 bits in eosio::name. This is why additional indices can exist up to 16. (4 bits, 0 - 15 will be translated into ., 1-5, a-j)

If you show the struct declarations and typedefs of mcat and mcatmig, I can answer more accurately.