À propos de cet exercice
Un jeu de données de 300 lignes sur une feuille, 3 cellules de réponse, environ 10 minutes. Corrigé dans votre navigateur : la valeur et la formule qui la produit, avec des indications sur la correction à apporter.
Comment le résoudre
- G2
=SUMIFS($D$2:$D$301,$C$2:$C$301,$F2,$B$2:$B$301,">="&G$1,$B$2:$B$301,"<"&EDATE(G$1,1)) - M2
=SUM(G2:L2) - G7
=SUM(G2:G6)
A PivotTable that groups dates by Months puts every order into the calendar month it falls in, from the 1st to the last day. The SUMIFS rebuilds that bucket with two conditions on the same date column — ">="&G$1 (on or after the month's 1st) and "<"&EDATE(G$1,1) (before the next month's 1st) — plus the category condition. Mixed references let one formula fill all 30 cells, and the grand totals reconcile to SUM(D2:D301) because every order lands in exactly one cell.
The PivotTable route in Excel: Insert → PivotTable, Category to Rows, Order date to Columns, Revenue to Values (Sum). Excel 2016 and later group a date field automatically; to set the grouping yourself, right-click a date → Group → Months (the labels then read Jan, Feb, …; select Years as well when the data spans more than one year, or January of both years lands in one column). The pivot needs Refresh after new data; these formulas recalculate by themselves, as long as a new export fits in rows 2–301 (for a longer one, extend the ranges or keep the data in an Excel Table). Pitfall 1: the bounds. ">"&G$1 drops every order of the 1st (the deal day), "<="&EDATE(G$1,1) counts the next month's 1st in two months, "<"&EOMONTH(G$1,0) drops the last day (the B2B invoices). Pitfall 2: timestamps. An order at 31 Jan 14:30 is later than EOMONTH(G$1,0), which is midnight, so "<="&EOMONTH(…) misses it, while "<"&EDATE(G$1,1) is safe with times too. Pitfall 3: text headers such as "Jan" cannot be compared with dates; with real dates in row 1 the same formula works for any month and year, and changing G1 re-targets its whole column. In Excel 365, =SUMIFS(D2:D301,C2:C301,F2:F6,B2:B301,">="&G1:L1,B2:B301,"<"&EDATE(+G1:L1,1)) in G2 spills the whole grid from one cell — the + in front of G1:L1 matters, because Excel's EDATE rejects a plain range with #VALUE! but accepts an array. PIVOTBY builds a cross-tab with its own headers (=PIVOTBY(C2:C301,TEXT(B2:B301,"yyyy-mm"),D2:D301,SUM)); this practice grid does not support PIVOTBY.
Autres méthodes valables
- =SUMIFS($D$2:$D$301,$C$2:$C$301,$F2,$B$2:$B$301,">="&G$1,$B$2:$B$301,"<="&EOMONTH(G$1,0)) — the month end as the upper bound (fine while the dates carry no times)
- =SUMPRODUCT(($C$2:$C$301=$F2)*(TEXT($B$2:$B$301,"yyyy-mm")=TEXT(G$1,"yyyy-mm"))*$D$2:$D$301) — compare year and month as text
- =SUMIFS(D2:D301,C2:C301,F2:F6,B2:B301,">="&G1:L1,B2:B301,"<"&EDATE(+G1:L1,1)) in G2 — Excel 365, one formula spills the whole grid (keep the + before G1:L1)
- =SUMIFS($D$2:$D$301,$C$2:$C$301,$F2) in M2 — a category total straight from the data instead of adding up the row
Erreurs fréquentes
Writing ">"&G$1 instead of ">="&G$1: every order placed on the 1st — the monthly deal day — drops out of its month, so G2 already misses part of the January Accessories revenue.
Comment nous corrigeons
Toute formule qui produit le bon résultat est acceptée, sauf si la tâche nomme une fonction : une autre fonction rapporte alors la moitié des points. Nous comparons les valeurs avec une petite tolérance et ignorons la casse et les espaces dans le texte. Un nombre correct saisi à la main au lieu d’une formule rapporte la moitié des points. Les tâches sur plusieurs cellules rapportent tous les points avec 100 % de cellules justes, la moitié à partir de 50 %. Lorsque c’est pertinent, nous modifions aussi une cellule d’entrée sur une copie de votre feuille et vérifions que votre formule suit.
Réponses calculées selon les règles d’Excel 365 et vérifiées dans notre correcteur le 20 novembre 2026. Auteur : GridBelt team.