#include <bits/stdc++.h>

using namespace std;
using ll = long long;
using VI = vector<int>;
using VL = vector<ll>;
using VS = vector<string>;
using VB = vector<bool>;
using VVI = vector<VI>;
using VVL = vector<VL>;
using PII = std::pair<int, int>;
using PLL = std::pair<ll, ll>;
using TI3 = std::tuple<int, int, int>;
using TI4 = std::tuple<int, int, int, int>;
using TL3 = std::tuple<ll, ll, ll>;
using TL4 = std::tuple<ll, ll, ll, ll>;

#define rep(i, n) for (int i = 0; i < (int)(n); i++)
#define repr(i, n) for (int i = (int)(n)-1; i >= 0; i--)
#define rep2(i, s, n) for (int i = (s); i < (int)(n); i++)
#define rep3(i, s, n, d) for (int i = (s); i < (int)(n); i += (d))
#define allpt(v) (v).begin(), (v).end()
#define allpt_c(v) (v).cbegin(), (v).cend()
#define allpt_r(v) (v).rbegin(), (v).rend()

const int mod1 = 1e9 + 7, mod2 = 998244353, mod3 = 1e9 + 9;
const int mod = mod1;
const ll inf = 1e18;

const string wsp = " ";
const string tb = "\t";
const string rt = "\n";

template <typename T>
void show1dvec(const vector<T> &v)
{
    if (v.size() == 0)
        return;
    int n = v.size() - 1;
    rep(i, n) cout << v[i] << wsp;
    cout << v[n] << rt;

    return;
}

template <typename T>
void show2dvec(const vector<vector<T>> &v)
{
    int n = v.size();
    rep(i, n) show1dvec(v[i]);
}

template <typename T, typename S>
void show1dpair(const vector<pair<T, S>> &v)
{
    int n = v.size();
    rep(i, n) cout << v[i].first << wsp << v[i].second << rt;
    return;
}

template <typename T, typename S>
void pairzip(const vector<pair<T, S>> &v, vector<T> &t, vector<T> &s)
{
    int n = v.size();
    rep(i, n)
    {
        t.push_back(v[i].first);
        s.push_back(v[i].second);
    }
    return;
}

template <typename T>
void maxvec(vector<T> &v)
{
    T s = v[0];
    int n = v.size();
    rep(i, n - 1)
    {
        if (s > v[i + 1])
        {
            v[i + 1] = s;
        }
        s = v[i + 1];
    }
}

template <typename T, typename S>
bool myfind(T t, S s)
{
    return find(t.cbegin(), t.cend(), s) != t.cend();
}

bool check(int y, int x, int h, int w)
{
    return 0 <= y && y < h && 0 <= x && x < w;
}

template <typename T>
vector<T> cumsum(const vector<T> &v)
{
    T s = 0;
    vector<T> ret;

    rep(i, v.size())
    {
        s += v[i];
        s %= mod;
        ret.emplace_back(s);
    }
    return ret;
}

bool iskadomatsu(int a, int b, int c)
{
    return (a != b && b != c && c != a) && ((a > b && b < c) || (a < b && b > c));
}

VS split(string s, char c)
{
    VS ret;
    string part;
    s += c;
    rep(i, s.length())
    {
        if (s[i] == c)
        {
            ret.emplace_back(part);
            part = "";
        }
        else if (s[i] != c)
        {
            part += s[i];
        }
    }
    return ret;
}

string removezero(string &s)
{
    string ret;
    for (auto z : s)
        if (z != '0')
            ret += z;
    return ret;
}

ll sumdigit(ll x)
{
    ll ans{0};
    while (x > 0)
    {
        ans += x % 10;
        x /= 10;
    }
    return ans;
}

template <typename T, typename S, typename R>
ll pow_mod(T p, S q, R mod = 1ll)
{
    ll ret = 1, r = p;
    while (q)
    {
        if (q % 2)
            ret *= r, ret %= mod;
        r = (r * r) % mod, q /= 2;
    }
    return ret % mod;
}

template <typename T, typename S>
ll pow_no_mod(T p, S q)
{
    ll ret = 1, r = p;
    while (q)
    {
        if (q % 2)
            ret *= r;
        r = (r * r), q /= 2;
    }
    return ret;
}

void make_frac_tables(VL &frac_list, VL &frac_inv_list)
{
    rep(i, frac_list.size() - 1)
    {
        frac_list[i + 1] *= frac_list[i] * (i + 1);
        frac_list[i + 1] %= mod;
        frac_inv_list[i + 1] *= frac_inv_list[i] * pow_mod(i + 1, mod - 2, mod);
        frac_inv_list[i + 1] %= mod;
    }
}

ll comb(int a, int b, const VL &frac_list, const VL &frac_inv_list)
{
    if (a < b)
        return 0;
    ll ret = frac_list[a];
    ret *= frac_inv_list[b];
    ret %= mod;
    ret *= frac_inv_list[a - b];
    ret %= mod;
    return ret;
}

class Matrix_n
{
    int matsize;
    VVL mat;

public:
    Matrix_n();
    Matrix_n(int n);
    Matrix_n(int n, const VVL &inmat);

    void showmat();
    void identify();
    ll getone(const int &i, const int &j);
    const Matrix_n operator+(const Matrix_n &b);
    const Matrix_n operator-(const Matrix_n &b);
    const Matrix_n operator*(const Matrix_n &b);
    Matrix_n &operator+=(const Matrix_n &b);
    Matrix_n &operator-=(const Matrix_n &b);
    const Matrix_n operator+();
    const Matrix_n operator-();
    const Matrix_n operator%(const int &b);
    Matrix_n &operator%=(const int &b);
};

Matrix_n::Matrix_n()
{
    matsize = 0;
    VVL mat;
}

Matrix_n::Matrix_n(int n)
{
    matsize = n;
    mat.resize(n);
    rep(i, n) rep(j, n)
        mat[i]
            .emplace_back(0);
}

Matrix_n::Matrix_n(int n, const VVL &inmat)
{
    matsize = n;
    mat.resize(n);
    rep(i, n)
    {
        mat[i].resize(n);
        rep(j, n)
        {
            mat[i][j] = inmat[i][j];
        }
    }
}

void Matrix_n::identify()
{
    rep(i, matsize) rep(j, matsize)
    {
        mat[i][j] = i == j ? 1 : 0;
    }
}

ll Matrix_n::getone(const int &i, const int &j)
{
    return mat[i][j];
}

const Matrix_n Matrix_n::operator+(const Matrix_n &b)
{
    Matrix_n ret(matsize);
    rep(i, matsize) rep(j, matsize)
    {
        ret.mat[i][j] = mat[i][j] + b.mat[i][j];
    }
    return ret;
}

const Matrix_n Matrix_n::operator-(const Matrix_n &b)
{
    Matrix_n ret(matsize);
    rep(i, matsize) rep(j, matsize)
    {
        ret.mat[i][j] = mat[i][j] - b.mat[i][j];
    }
    return ret;
}

const Matrix_n Matrix_n::operator*(const Matrix_n &b)
{
    Matrix_n ret(matsize);
    rep(i, matsize) rep(j, matsize) rep(k, matsize)
    {
        ret.mat[i][j] += mat[i][k] * b.mat[k][j];
    }
    return ret;
}

const Matrix_n Matrix_n::operator%(const int &b)
{
    Matrix_n ret(matsize);
    rep(i, matsize) rep(j, matsize)
    {
        ret.mat[i][j] = mat[i][j] % b;
    }
    return ret;
}

const Matrix_n Matrix_n::operator+()
{
    return *this;
}

const Matrix_n Matrix_n::operator-()
{
    Matrix_n ret(matsize);
    rep(i, matsize) rep(j, matsize)
    {
        ret.mat[i][j] = -mat[i][j];
    }
    return ret;
}

Matrix_n &Matrix_n::operator+=(const Matrix_n &b)
{
    rep(i, matsize) rep(j, matsize)
    {
        mat[i][j] += b.mat[i][j];
    }
    return *this;
}

Matrix_n &Matrix_n::operator-=(const Matrix_n &b)
{
    rep(i, matsize) rep(j, matsize)
    {
        mat[i][j] -= b.mat[i][j];
    }
    return *this;
}

Matrix_n &Matrix_n::operator%=(const int &b)
{
    rep(i, matsize) rep(j, matsize)
    {
        mat[i][j] %= b;
    }
    return *this;
}

void Matrix_n::showmat()
{
    show2dvec(mat);
}


int main()
{

    cin.tie(0);
    ios::sync_with_stdio(false);

#ifdef DEBUG
    cout << "DEBUG MODE" << endl;
    ifstream in("input.txt"); //for debug
    cin.rdbuf(in.rdbuf());    //for debug
#endif

    int n;
    cin >> n;
    VI num(n), ans(n), from_index;
    iota(allpt(ans), 0);
    deque<int> to_index;
    rep(i, n) cin >> num[i];
    rep(i, n)
    {
        if (num[i] == i)
        {
            from_index.emplace_back(i);
            to_index.emplace_back(i);
        }
    }

    if (from_index.empty())
    {
        rep(i, n)
        {
            cout << ans[i] << rt;
        }
    }

    else if (from_index.size() > 1)
    {
        to_index.emplace_front(to_index.back());
        to_index.pop_back();
        rep(i, from_index.size())
        {
            ans[from_index[i]] = to_index[i];
        }
        rep(i, n)
        {
            cout << ans[i] << rt;
        }
    }

    else
    {
        int a = from_index[0];
        if (count(allpt_c(num), a) == n)
        {
            cout << -1 << rt;
        }
        else
        {
            rep(i, n)
            {
                if (num[i] != a)
                {
                    swap(ans[a], ans[i]);
                    break;
                }
            }
            rep(i, n)
            {
                cout << ans[i] << rt;
            }
        }
    }

    return 0;
}