Re: MonetDB: viewless - remove BATslice of candidate lists from gdk_...
No idea what viewless branch is aiming at, but my 5 1/2th sense tells me there is something interesting here.. ;)
On 27 Apr 2017 5:56 pm, "lsidir" commits+lsidir=gmail.com@monetdb.org wrote:
Changeset: 63c8d8dbed74 for MonetDB URL: http://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=63c8d8dbed74 Modified Files: gdk/gdk_firstn.c Branch: viewless Log Message:
remove BATslice of candidate lists from gdk_firstn
diffs (45 lines):
diff --git a/gdk/gdk_firstn.c b/gdk/gdk_firstn.c --- a/gdk/gdk_firstn.c +++ b/gdk/gdk_firstn.c @@ -140,7 +140,7 @@ BATfirstn_unique(BAT *b, BAT *s, BUN n, if (n >= (BUN) (candend - cand)) { /* trivial: return the candidate list (the * part that refers to b, that is) */
return BATslice(s,
return CANDslice(s, (BUN) (cand - (const oid *)Tloc(s, 0)), (BUN) (candend - (const oid *) Tloc(s, 0))); } @@ -163,12 +163,12 @@ BATfirstn_unique(BAT *b, BAT *s, BUN n, /* return copy of first relevant part * of candidate list */ i = (BUN) (cand - (const oid *) Tloc(s, 0));
return BATslice(s, i, i + n);
return CANDslice(s, i, i + n); } /* return copy of last relevant part of * candidate list */ i = (BUN) (candend - (const oid *) Tloc(s, 0));
return BATslice(s, i - n, i);
return CANDslice(s, i - n, i); } bn = COLnew(0, TYPE_void, n, TRANSIENT); if (bn == NULL)@@ -374,7 +374,7 @@ BATfirstn_unique_with_groups(BAT *b, BAT if (BATtdense(g)) { /* trivial: g determines ordering, return initial * slice of s */
return BATslice(s, 0, n);
return CANDslice(s, 0, n); } CANDINIT(b, s, start, end, cnt, cand, candend);@@ -916,7 +916,7 @@ BATfirstn_grouped_with_groups(BAT **topn if (BATtdense(g)) { /* trivial: g determines ordering, return initial * slice of s */
bn = BATslice(s, 0, n);
bn = CANDslice(s, 0, n); gn = gids ? BATslice(g, 0, n) : NULL; if (bn == NULL || (gids != NULL && gn == NULL)) { BBPreclaim(bn);
checkin-list mailing list checkin-list@monetdb.org https://www.monetdb.org/mailman/listinfo/checkin-list
I am trying to get rid of all the views that are defined over a bat. Basically, we will not have parentid anymore, thus no bats that point to a slice of a heap of another bat. This last one will only be possible with a dense candidate list over the parent (and hence the only) bat. Almost all gdk functions support candidate lists, so this will be easy to achieve.
The benefit for this is cleaner code, since we will not have to check if a bat is a view, and do the sometimes weird positional math. But most importantly, indexes that are defined on the parent bat (which is usually also a persistent one) will be easier to use. Up until now, in order to use an index on a view, the index should be "slice-able" too. Now, we only need to augment the index access with candidate lists. The goal is to lead to more often use of the persistent indexes than what we do up until now (which access is usually killed by the slicing of mitosis).
It was good that i wrote it down so i also have a record of the goal:)
cheers,
lefteris
On Thu, Apr 27, 2017 at 6:25 PM, Roberto Cornacchia roberto.cornacchia@gmail.com wrote:
No idea what viewless branch is aiming at, but my 5 1/2th sense tells me there is something interesting here.. ;)
On 27 Apr 2017 5:56 pm, "lsidir" commits+lsidir=gmail.com@monetdb.org wrote:
Changeset: 63c8d8dbed74 for MonetDB URL: http://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=63c8d8dbed74 Modified Files: gdk/gdk_firstn.c Branch: viewless Log Message:
remove BATslice of candidate lists from gdk_firstn
diffs (45 lines):
diff --git a/gdk/gdk_firstn.c b/gdk/gdk_firstn.c --- a/gdk/gdk_firstn.c +++ b/gdk/gdk_firstn.c @@ -140,7 +140,7 @@ BATfirstn_unique(BAT *b, BAT *s, BUN n, if (n >= (BUN) (candend - cand)) { /* trivial: return the candidate list (the * part that refers to b, that is) */
return BATslice(s,
return CANDslice(s, (BUN) (cand - (const oid *)Tloc(s, 0)), (BUN) (candend - (const oid *) Tloc(s, 0))); } @@ -163,12 +163,12 @@ BATfirstn_unique(BAT *b, BAT *s, BUN n, /* return copy of first relevant part * of candidate list */ i = (BUN) (cand - (const oid *) Tloc(s, 0));
return BATslice(s, i, i + n);
return CANDslice(s, i, i + n); } /* return copy of last relevant part of * candidate list */ i = (BUN) (candend - (const oid *) Tloc(s, 0));
return BATslice(s, i - n, i);
return CANDslice(s, i - n, i); } bn = COLnew(0, TYPE_void, n, TRANSIENT); if (bn == NULL)@@ -374,7 +374,7 @@ BATfirstn_unique_with_groups(BAT *b, BAT if (BATtdense(g)) { /* trivial: g determines ordering, return initial * slice of s */
return BATslice(s, 0, n);
return CANDslice(s, 0, n); } CANDINIT(b, s, start, end, cnt, cand, candend);@@ -916,7 +916,7 @@ BATfirstn_grouped_with_groups(BAT **topn if (BATtdense(g)) { /* trivial: g determines ordering, return initial * slice of s */
bn = BATslice(s, 0, n);
bn = CANDslice(s, 0, n); gn = gids ? BATslice(g, 0, n) : NULL; if (bn == NULL || (gids != NULL && gn == NULL)) { BBPreclaim(bn);
checkin-list mailing list checkin-list@monetdb.org https://www.monetdb.org/mailman/listinfo/checkin-list
developers-list mailing list developers-list@monetdb.org https://www.monetdb.org/mailman/listinfo/developers-list
Sounds great, thanks explaining! Cheers, Roberto
On 28 April 2017 at 10:06, Lefteris lsidir@gmail.com wrote:
I am trying to get rid of all the views that are defined over a bat. Basically, we will not have parentid anymore, thus no bats that point to a slice of a heap of another bat. This last one will only be possible with a dense candidate list over the parent (and hence the only) bat. Almost all gdk functions support candidate lists, so this will be easy to achieve.
The benefit for this is cleaner code, since we will not have to check if a bat is a view, and do the sometimes weird positional math. But most importantly, indexes that are defined on the parent bat (which is usually also a persistent one) will be easier to use. Up until now, in order to use an index on a view, the index should be "slice-able" too. Now, we only need to augment the index access with candidate lists. The goal is to lead to more often use of the persistent indexes than what we do up until now (which access is usually killed by the slicing of mitosis).
It was good that i wrote it down so i also have a record of the goal:)
cheers,
lefteris
On Thu, Apr 27, 2017 at 6:25 PM, Roberto Cornacchia roberto.cornacchia@gmail.com wrote:
No idea what viewless branch is aiming at, but my 5 1/2th sense tells me there is something interesting here.. ;)
On 27 Apr 2017 5:56 pm, "lsidir" commits+lsidir=gmail.com@monetdb.org wrote:
Changeset: 63c8d8dbed74 for MonetDB URL: http://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=63c8d8dbed74 Modified Files: gdk/gdk_firstn.c Branch: viewless Log Message:
remove BATslice of candidate lists from gdk_firstn
diffs (45 lines):
diff --git a/gdk/gdk_firstn.c b/gdk/gdk_firstn.c --- a/gdk/gdk_firstn.c +++ b/gdk/gdk_firstn.c @@ -140,7 +140,7 @@ BATfirstn_unique(BAT *b, BAT *s, BUN n, if (n >= (BUN) (candend - cand)) { /* trivial: return the candidate list (the * part that refers to b, that is) */
return BATslice(s,
return CANDslice(s, (BUN) (cand - (const oid *)Tloc(s, 0)), (BUN) (candend - (const oid *) Tloc(s, 0))); } @@ -163,12 +163,12 @@ BATfirstn_unique(BAT *b, BAT *s, BUN n, /* return copy of first relevant part * of candidate list */ i = (BUN) (cand - (const oid *) Tloc(s, 0));
return BATslice(s, i, i + n);
return CANDslice(s, i, i + n); } /* return copy of last relevant part of * candidate list */ i = (BUN) (candend - (const oid *) Tloc(s, 0));
return BATslice(s, i - n, i);
return CANDslice(s, i - n, i); } bn = COLnew(0, TYPE_void, n, TRANSIENT); if (bn == NULL)@@ -374,7 +374,7 @@ BATfirstn_unique_with_groups(BAT *b, BAT if (BATtdense(g)) { /* trivial: g determines ordering, return initial * slice of s */
return BATslice(s, 0, n);
return CANDslice(s, 0, n); } CANDINIT(b, s, start, end, cnt, cand, candend);@@ -916,7 +916,7 @@ BATfirstn_grouped_with_groups(BAT **topn if (BATtdense(g)) { /* trivial: g determines ordering, return initial * slice of s */
bn = BATslice(s, 0, n);
bn = CANDslice(s, 0, n); gn = gids ? BATslice(g, 0, n) : NULL; if (bn == NULL || (gids != NULL && gn == NULL)) { BBPreclaim(bn);
checkin-list mailing list checkin-list@monetdb.org https://www.monetdb.org/mailman/listinfo/checkin-list
developers-list mailing list developers-list@monetdb.org https://www.monetdb.org/mailman/listinfo/developers-list
developers-list mailing list developers-list@monetdb.org https://www.monetdb.org/mailman/listinfo/developers-list
participants (2)
-
Lefteris -
Roberto Cornacchia